Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* GStreamer Matroska muxer/demuxer
|
|
|
|
* (c) 2003 Ronald Bultje <rbultje@ronald.bitfreak.net>
|
2005-10-14 12:43:30 +00:00
|
|
|
* (c) 2005 Michal Benes <michal.benes@xeris.cz>
|
2008-08-02 17:47:32 +00:00
|
|
|
* (c) 2008 Sebastian Dröge <sebastian.droege@collabora.co.uk>
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
*
|
|
|
|
* matroska-mux.c: matroska file/stream muxer
|
|
|
|
*
|
|
|
|
* This library is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU Library General Public
|
|
|
|
* License as published by the Free Software Foundation; either
|
|
|
|
* version 2 of the License, or (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This library is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
* Library General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU Library General Public
|
|
|
|
* License along with this library; if not, write to the
|
|
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
|
|
* Boston, MA 02111-1307, USA.
|
|
|
|
*/
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: - check everywhere that we don't write invalid values
|
|
|
|
* - make sure timestamps are correctly scaled everywhere
|
|
|
|
*/
|
|
|
|
|
2008-12-05 09:24:18 +00:00
|
|
|
/**
|
|
|
|
* SECTION:element-matroskamux
|
|
|
|
*
|
|
|
|
* matroskamux muxes different input streams into a Matroska file.
|
|
|
|
*
|
|
|
|
* <refsect2>
|
|
|
|
* <title>Example launch line</title>
|
|
|
|
* |[
|
|
|
|
* gst-launch -v filesrc location=/path/to/mp3 ! mp3parse ! matroskamux name=mux ! filesink location=test.mkv filesrc location=/path/to/theora.ogg ! oggdemux ! theoraparse ! mux.
|
|
|
|
* ]| This pipeline muxes an MP3 file and a Ogg Theora video into a Matroska file.
|
|
|
|
* |[
|
|
|
|
* gst-launch -v audiotestsrc num-buffers=100 ! audioconvert ! vorbisenc ! matroskamux ! filesink location=test.mka
|
|
|
|
* ]| This pipeline muxes a 440Hz sine wave encoded with the Vorbis codec into a Matroska file.
|
|
|
|
* </refsect2>
|
|
|
|
*/
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
#ifdef HAVE_CONFIG_H
|
|
|
|
#include "config.h"
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#include <math.h>
|
|
|
|
#include <string.h>
|
|
|
|
|
2009-08-08 03:47:48 +00:00
|
|
|
#include <gst/riff/riff-media.h>
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
#include "matroska-mux.h"
|
|
|
|
#include "matroska-ids.h"
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
GST_DEBUG_CATEGORY_STATIC (matroskamux_debug);
|
2004-07-18 15:58:04 +00:00
|
|
|
#define GST_CAT_DEFAULT matroskamux_debug
|
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
enum
|
|
|
|
{
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
ARG_0,
|
2005-10-28 15:32:48 +00:00
|
|
|
ARG_WRITING_APP,
|
|
|
|
ARG_MATROSKA_VERSION
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
};
|
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
#define DEFAULT_MATROSKA_VERSION 1
|
|
|
|
#define DEFAULT_WRITING_APP "GStreamer Matroska muxer"
|
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
static GstStaticPadTemplate src_templ = GST_STATIC_PAD_TEMPLATE ("src",
|
|
|
|
GST_PAD_SRC,
|
|
|
|
GST_PAD_ALWAYS,
|
|
|
|
GST_STATIC_CAPS ("video/x-matroska")
|
|
|
|
);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
#define COMMON_VIDEO_CAPS \
|
|
|
|
"width = (int) [ 16, 4096 ], " \
|
|
|
|
"height = (int) [ 16, 4096 ], " \
|
2005-11-22 22:21:37 +00:00
|
|
|
"framerate = (fraction) [ 0, MAX ]"
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2008-01-08 20:03:30 +00:00
|
|
|
#define COMMON_VIDEO_CAPS_NO_FRAMERATE \
|
|
|
|
"width = (int) [ 16, 4096 ], " \
|
|
|
|
"height = (int) [ 16, 4096 ] "
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* FIXME:
|
|
|
|
* * require codec data, etc as needed
|
|
|
|
*/
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
static GstStaticPadTemplate videosink_templ =
|
2004-03-14 22:34:33 +00:00
|
|
|
GST_STATIC_PAD_TEMPLATE ("video_%d",
|
|
|
|
GST_PAD_SINK,
|
|
|
|
GST_PAD_REQUEST,
|
|
|
|
GST_STATIC_CAPS ("video/mpeg, "
|
2004-03-15 19:32:27 +00:00
|
|
|
"mpegversion = (int) { 1, 2, 4 }, "
|
|
|
|
"systemstream = (boolean) false, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2005-10-14 12:43:30 +00:00
|
|
|
"video/x-h264, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2004-03-15 19:32:27 +00:00
|
|
|
"video/x-divx, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
|
|
|
"video/x-xvid, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2006-04-25 11:45:00 +00:00
|
|
|
"video/x-huffyuv, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
|
|
|
"video/x-dv, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
|
|
|
"video/x-h263, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2004-03-15 19:32:27 +00:00
|
|
|
"video/x-msmpeg, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2005-10-18 18:06:14 +00:00
|
|
|
"image/jpeg, "
|
2008-01-08 20:03:30 +00:00
|
|
|
COMMON_VIDEO_CAPS_NO_FRAMERATE "; "
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
"video/x-theora; "
|
2006-11-15 00:12:19 +00:00
|
|
|
"video/x-dirac, "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2008-08-25 14:15:43 +00:00
|
|
|
"video/x-pn-realvideo, "
|
|
|
|
"rmversion = (int) [1, 4], "
|
|
|
|
COMMON_VIDEO_CAPS "; "
|
2004-03-15 19:32:27 +00:00
|
|
|
"video/x-raw-yuv, "
|
2008-08-02 17:54:04 +00:00
|
|
|
"format = (fourcc) { YUY2, I420, YV12, UYVY, AYUV }, "
|
2009-08-06 23:15:17 +00:00
|
|
|
COMMON_VIDEO_CAPS "; "
|
|
|
|
"video/x-wmv, " "wmvversion = (int) [ 1, 3 ], " COMMON_VIDEO_CAPS)
|
2004-03-14 22:34:33 +00:00
|
|
|
);
|
2003-12-22 01:47:09 +00:00
|
|
|
|
|
|
|
#define COMMON_AUDIO_CAPS \
|
2008-08-02 17:52:16 +00:00
|
|
|
"channels = (int) [ 1, MAX ], " \
|
|
|
|
"rate = (int) [ 1, MAX ]"
|
configure.ac: Improve mpeg2enc detection. This is for distributions that do ship mjpegtools, but without mpeg2enc. Al...
Original commit message from CVS:
2003-12-21 Ronald Bultje <rbultje@ronald.bitfreak.net>
* configure.ac:
Improve mpeg2enc detection. This is for distributions that do
ship mjpegtools, but without mpeg2enc. Also does object check
for might there ever be ABI incompatibility.
* ext/mpeg2enc/gstmpeg2enc.cc:
Add Andrew as second maintainer (he's helping me), and also add
an error if no caps was set. This happens if I pull before capsnego
and that's something I should solve sometime else.
* gst/matroska/matroska-demux.c:
(gst_matroska_demux_parse_blockgroup):
Fix time parsing.
* gst/matroska/matroska-mux.c: (gst_matroska_mux_audio_pad_link),
(gst_matroska_mux_track_header):
Add caps to templates.
* gst/mpegaudioparse/gstmpegaudioparse.c: (mp3_sink_factory):
Add mpegversion=1 to prevent confusion with MPEG/AAC.
* gst/mpegstream/gstmpegdemux.c:
Remove layer since it causes warnings about unfixed caps.
* gst/videotestsrc/gstvideotestsrc.c: (gst_videotestsrc_get):
Fix obvious typo (we error out if caps were set, we should of
course error out if *no* caps were set).
* sys/oss/gstosselement.c: (gst_osselement_convert):
Fix format conversion, we confused bits/bytes.
* sys/oss/gstosselement.h:
Improve documentation for 'bps'.
* sys/v4l/TODO:
Remove stuff about plugins that need removing - this was done
ages ago.
* sys/v4l/gstv4lmjpegsrc.c: (gst_v4lmjpegsrc_init),
(gst_v4lmjpegsrc_src_convert), (gst_v4lmjpegsrc_src_query):
* sys/v4l/gstv4lsrc.c: (gst_v4lsrc_init), (gst_v4lsrc_src_convert),
(gst_v4lsrc_src_query):
* sys/v4l2/gstv4l2src.c: (gst_v4l2src_init),
(gst_v4l2src_src_convert), (gst_v4l2src_src_query):
Add get_query_types(), get_formats() and query() functions.
2003-12-21 18:44:34 +00:00
|
|
|
|
|
|
|
/* FIXME:
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
* * require codec data, etc as needed
|
configure.ac: Improve mpeg2enc detection. This is for distributions that do ship mjpegtools, but without mpeg2enc. Al...
Original commit message from CVS:
2003-12-21 Ronald Bultje <rbultje@ronald.bitfreak.net>
* configure.ac:
Improve mpeg2enc detection. This is for distributions that do
ship mjpegtools, but without mpeg2enc. Also does object check
for might there ever be ABI incompatibility.
* ext/mpeg2enc/gstmpeg2enc.cc:
Add Andrew as second maintainer (he's helping me), and also add
an error if no caps was set. This happens if I pull before capsnego
and that's something I should solve sometime else.
* gst/matroska/matroska-demux.c:
(gst_matroska_demux_parse_blockgroup):
Fix time parsing.
* gst/matroska/matroska-mux.c: (gst_matroska_mux_audio_pad_link),
(gst_matroska_mux_track_header):
Add caps to templates.
* gst/mpegaudioparse/gstmpegaudioparse.c: (mp3_sink_factory):
Add mpegversion=1 to prevent confusion with MPEG/AAC.
* gst/mpegstream/gstmpegdemux.c:
Remove layer since it causes warnings about unfixed caps.
* gst/videotestsrc/gstvideotestsrc.c: (gst_videotestsrc_get):
Fix obvious typo (we error out if caps were set, we should of
course error out if *no* caps were set).
* sys/oss/gstosselement.c: (gst_osselement_convert):
Fix format conversion, we confused bits/bytes.
* sys/oss/gstosselement.h:
Improve documentation for 'bps'.
* sys/v4l/TODO:
Remove stuff about plugins that need removing - this was done
ages ago.
* sys/v4l/gstv4lmjpegsrc.c: (gst_v4lmjpegsrc_init),
(gst_v4lmjpegsrc_src_convert), (gst_v4lmjpegsrc_src_query):
* sys/v4l/gstv4lsrc.c: (gst_v4lsrc_init), (gst_v4lsrc_src_convert),
(gst_v4lsrc_src_query):
* sys/v4l2/gstv4l2src.c: (gst_v4l2src_init),
(gst_v4l2src_src_convert), (gst_v4l2src_src_query):
Add get_query_types(), get_formats() and query() functions.
2003-12-21 18:44:34 +00:00
|
|
|
*/
|
2003-12-22 01:47:09 +00:00
|
|
|
static GstStaticPadTemplate audiosink_templ =
|
2004-03-14 22:34:33 +00:00
|
|
|
GST_STATIC_PAD_TEMPLATE ("audio_%d",
|
|
|
|
GST_PAD_SINK,
|
|
|
|
GST_PAD_REQUEST,
|
|
|
|
GST_STATIC_CAPS ("audio/mpeg, "
|
2004-03-15 19:32:27 +00:00
|
|
|
"mpegversion = (int) 1, "
|
|
|
|
"layer = (int) [ 1, 3 ], "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
|
|
|
"audio/mpeg, "
|
|
|
|
"mpegversion = (int) { 2, 4 }, "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
|
|
|
"audio/x-ac3, "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
2005-10-14 12:43:30 +00:00
|
|
|
"audio/x-vorbis, "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
2008-08-09 13:48:22 +00:00
|
|
|
"audio/x-flac, "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
2009-04-13 12:03:03 +00:00
|
|
|
"audio/x-speex, "
|
|
|
|
COMMON_AUDIO_CAPS "; "
|
2004-03-15 19:32:27 +00:00
|
|
|
"audio/x-raw-int, "
|
2006-12-24 11:24:59 +00:00
|
|
|
"width = (int) 8, "
|
|
|
|
"depth = (int) 8, "
|
|
|
|
"signed = (boolean) false, "
|
|
|
|
COMMON_AUDIO_CAPS ";"
|
|
|
|
"audio/x-raw-int, "
|
|
|
|
"width = (int) 16, "
|
|
|
|
"depth = (int) 16, "
|
2004-03-15 19:32:27 +00:00
|
|
|
"endianness = (int) { BIG_ENDIAN, LITTLE_ENDIAN }, "
|
2006-12-24 11:24:59 +00:00
|
|
|
"signed = (boolean) true, "
|
2004-08-26 11:10:16 +00:00
|
|
|
COMMON_AUDIO_CAPS ";"
|
2008-08-02 17:52:16 +00:00
|
|
|
"audio/x-raw-int, "
|
|
|
|
"width = (int) 24, "
|
|
|
|
"depth = (int) 24, "
|
|
|
|
"endianness = (int) { BIG_ENDIAN, LITTLE_ENDIAN }, "
|
|
|
|
"signed = (boolean) true, "
|
|
|
|
COMMON_AUDIO_CAPS ";"
|
|
|
|
"audio/x-raw-int, "
|
|
|
|
"width = (int) 32, "
|
|
|
|
"depth = (int) 32, "
|
|
|
|
"endianness = (int) { BIG_ENDIAN, LITTLE_ENDIAN }, "
|
|
|
|
"signed = (boolean) true, "
|
|
|
|
COMMON_AUDIO_CAPS ";"
|
|
|
|
"audio/x-raw-float, "
|
|
|
|
"width = (int) [ 32, 64 ], "
|
|
|
|
"endianness = (int) LITTLE_ENDIAN, "
|
|
|
|
COMMON_AUDIO_CAPS ";"
|
2004-09-15 19:43:37 +00:00
|
|
|
"audio/x-tta, "
|
2004-08-26 11:10:16 +00:00
|
|
|
"width = (int) { 8, 16, 24 }, "
|
2008-08-25 14:15:43 +00:00
|
|
|
"channels = (int) { 1, 2 }, " "rate = (int) [ 8000, 96000 ]; "
|
|
|
|
"audio/x-pn-realaudio, "
|
2009-08-08 03:47:48 +00:00
|
|
|
"raversion = (int) { 1, 2, 8 }, " COMMON_AUDIO_CAPS "; "
|
|
|
|
"audio/x-wma, " "wmaversion = (int) [ 1, 3 ], "
|
|
|
|
"block_align = (int) [ 0, 65535 ], bitrate = (int) [ 0, 524288 ], "
|
|
|
|
COMMON_AUDIO_CAPS)
|
2004-03-14 22:34:33 +00:00
|
|
|
);
|
2003-12-22 01:47:09 +00:00
|
|
|
|
|
|
|
static GstStaticPadTemplate subtitlesink_templ =
|
2004-03-14 22:34:33 +00:00
|
|
|
GST_STATIC_PAD_TEMPLATE ("subtitle_%d",
|
|
|
|
GST_PAD_SINK,
|
|
|
|
GST_PAD_REQUEST,
|
|
|
|
GST_STATIC_CAPS_ANY);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2004-09-01 12:10:21 +00:00
|
|
|
static GArray *used_uids;
|
2008-08-02 17:47:32 +00:00
|
|
|
G_LOCK_DEFINE_STATIC (used_uids);
|
2004-09-01 12:10:21 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
static void gst_matroska_mux_add_interfaces (GType type);
|
|
|
|
|
|
|
|
GST_BOILERPLATE_FULL (GstMatroskaMux, gst_matroska_mux, GstElement,
|
|
|
|
GST_TYPE_ELEMENT, gst_matroska_mux_add_interfaces);
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/* Matroska muxer destructor */
|
|
|
|
static void gst_matroska_mux_finalize (GObject * object);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/* Pads collected callback */
|
|
|
|
static GstFlowReturn
|
|
|
|
gst_matroska_mux_collected (GstCollectPads * pads, gpointer user_data);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* pad functions */
|
2005-10-14 12:43:30 +00:00
|
|
|
static gboolean gst_matroska_mux_handle_src_event (GstPad * pad,
|
|
|
|
GstEvent * event);
|
2004-03-14 22:34:33 +00:00
|
|
|
static GstPad *gst_matroska_mux_request_new_pad (GstElement * element,
|
|
|
|
GstPadTemplate * templ, const gchar * name);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
static void gst_matroska_mux_release_pad (GstElement * element, GstPad * pad);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* gst internal change state handler */
|
2005-09-02 15:44:50 +00:00
|
|
|
static GstStateChangeReturn
|
|
|
|
gst_matroska_mux_change_state (GstElement * element, GstStateChange transition);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* gobject bla bla */
|
2004-03-14 22:34:33 +00:00
|
|
|
static void gst_matroska_mux_set_property (GObject * object,
|
|
|
|
guint prop_id, const GValue * value, GParamSpec * pspec);
|
|
|
|
static void gst_matroska_mux_get_property (GObject * object,
|
|
|
|
guint prop_id, GValue * value, GParamSpec * pspec);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* reset muxer */
|
2004-03-14 22:34:33 +00:00
|
|
|
static void gst_matroska_mux_reset (GstElement * element);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2004-09-01 12:10:21 +00:00
|
|
|
/* uid generation */
|
2008-08-02 17:47:32 +00:00
|
|
|
static guint64 gst_matroska_mux_create_uid ();
|
2004-09-01 12:10:21 +00:00
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
static gboolean theora_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context);
|
|
|
|
static gboolean vorbis_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context);
|
2009-04-13 12:03:03 +00:00
|
|
|
static gboolean speex_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context);
|
2009-02-15 18:49:44 +00:00
|
|
|
static gboolean kate_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context);
|
2008-08-09 13:48:22 +00:00
|
|
|
static gboolean flac_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
static void
|
|
|
|
gst_matroska_mux_add_interfaces (GType type)
|
|
|
|
{
|
|
|
|
static const GInterfaceInfo tag_setter_info = { NULL, NULL, NULL };
|
|
|
|
|
|
|
|
g_type_add_interface_static (type, GST_TYPE_TAG_SETTER, &tag_setter_info);
|
|
|
|
}
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static void
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_matroska_mux_base_init (gpointer g_class)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstElementClass *element_class = GST_ELEMENT_CLASS (g_class);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
gst_element_class_add_pad_template (element_class,
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_static_pad_template_get (&videosink_templ));
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_element_class_add_pad_template (element_class,
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_static_pad_template_get (&audiosink_templ));
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_element_class_add_pad_template (element_class,
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_static_pad_template_get (&subtitlesink_templ));
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_element_class_add_pad_template (element_class,
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_static_pad_template_get (&src_templ));
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
gst_element_class_set_details_simple (element_class, "Matroska muxer",
|
|
|
|
"Codec/Muxer",
|
|
|
|
"Muxes video/audio/subtitle streams into a matroska stream",
|
2009-08-24 20:42:42 +00:00
|
|
|
"GStreamer maintainers <gstreamer-devel@lists.sourceforge.net>");
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
|
|
|
|
GST_DEBUG_CATEGORY_INIT (matroskamux_debug, "matroskamux", 0,
|
|
|
|
"Matroska muxer");
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_class_init (GstMatroskaMuxClass * klass)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GObjectClass *gobject_class;
|
|
|
|
GstElementClass *gstelement_class;
|
|
|
|
|
|
|
|
gobject_class = (GObjectClass *) klass;
|
|
|
|
gstelement_class = (GstElementClass *) klass;
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
gobject_class->finalize = gst_matroska_mux_finalize;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
gobject_class->get_property = gst_matroska_mux_get_property;
|
|
|
|
gobject_class->set_property = gst_matroska_mux_set_property;
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
g_object_class_install_property (gobject_class, ARG_WRITING_APP,
|
|
|
|
g_param_spec_string ("writing-app", "Writing application.",
|
|
|
|
"The name the application that creates the matroska file.",
|
|
|
|
NULL, G_PARAM_READWRITE));
|
2005-10-28 15:32:48 +00:00
|
|
|
g_object_class_install_property (gobject_class, ARG_MATROSKA_VERSION,
|
|
|
|
g_param_spec_int ("version", "Matroska version",
|
|
|
|
"This parameter determines what matroska features can be used.",
|
2009-01-08 15:56:46 +00:00
|
|
|
1, 2, DEFAULT_MATROSKA_VERSION, G_PARAM_READWRITE));
|
2005-10-14 12:43:30 +00:00
|
|
|
|
2006-11-16 15:36:48 +00:00
|
|
|
gstelement_class->change_state =
|
|
|
|
GST_DEBUG_FUNCPTR (gst_matroska_mux_change_state);
|
|
|
|
gstelement_class->request_new_pad =
|
|
|
|
GST_DEBUG_FUNCPTR (gst_matroska_mux_request_new_pad);
|
|
|
|
gstelement_class->release_pad =
|
|
|
|
GST_DEBUG_FUNCPTR (gst_matroska_mux_release_pad);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_init:
|
|
|
|
* @mux: #GstMatroskaMux that should be initialized.
|
|
|
|
* @g_class: Class of the muxer.
|
|
|
|
*
|
|
|
|
* Matroska muxer constructor.
|
|
|
|
*/
|
2004-03-14 22:34:33 +00:00
|
|
|
static void
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_matroska_mux_init (GstMatroskaMux * mux, GstMatroskaMuxClass * g_class)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
mux->srcpad = gst_pad_new_from_static_template (&src_templ, "src");
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_pad_set_event_function (mux->srcpad, gst_matroska_mux_handle_src_event);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_element_add_pad (GST_ELEMENT (mux), mux->srcpad);
|
|
|
|
|
Fix for collect pads API change. Also fix textoverlay state change function.
Original commit message from CVS:
* ext/cairo/gsttextoverlay.c: (gst_text_overlay_finalize),
(gst_text_overlay_init), (gst_text_overlay_text_pad_linked),
(gst_text_overlay_text_pad_unlinked), (gst_text_overlay_pop_video),
(gst_text_overlay_pop_text), (gst_text_overlay_collected),
(gst_text_overlay_change_state):
* gst/matroska/matroska-mux.c: (gst_matroska_mux_init),
(gst_matroska_mux_reset), (gst_matroska_mux_request_new_pad),
(gst_matroska_mux_best_pad), (gst_matroska_mux_change_state):
* gst/smpte/gstsmpte.c: (gst_smpte_init), (gst_smpte_collected):
* gst/videomixer/videomixer.c: (gst_videomixer_init),
(gst_videomixer_request_new_pad), (gst_videomixer_fill_queues),
(gst_videomixer_change_state):
Fix for collect pads API change. Also fix textoverlay state
change function.
2005-11-21 12:13:48 +00:00
|
|
|
mux->collect = gst_collect_pads_new ();
|
|
|
|
gst_collect_pads_set_function (mux->collect,
|
2006-03-24 17:11:56 +00:00
|
|
|
(GstCollectPadsFunction) GST_DEBUG_FUNCPTR (gst_matroska_mux_collected),
|
|
|
|
mux);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
mux->ebml_write = gst_ebml_write_new (mux->srcpad);
|
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* property defaults */
|
|
|
|
mux->matroska_version = DEFAULT_MATROSKA_VERSION;
|
|
|
|
mux->writing_app = g_strdup (DEFAULT_WRITING_APP);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/* initialize internal variables */
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
mux->index = NULL;
|
2009-01-08 15:56:46 +00:00
|
|
|
mux->num_streams = 0;
|
|
|
|
mux->num_a_streams = 0;
|
|
|
|
mux->num_t_streams = 0;
|
|
|
|
mux->num_v_streams = 0;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* initialize remaining variables */
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_matroska_mux_reset (GST_ELEMENT (mux));
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_finalize:
|
|
|
|
* @object: #GstMatroskaMux that should be finalized.
|
|
|
|
*
|
|
|
|
* Finalize matroska muxer.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
gst_matroska_mux_finalize (GObject * object)
|
|
|
|
{
|
|
|
|
GstMatroskaMux *mux = GST_MATROSKA_MUX (object);
|
|
|
|
|
|
|
|
gst_object_unref (mux->collect);
|
|
|
|
gst_object_unref (mux->ebml_write);
|
2005-11-01 15:11:16 +00:00
|
|
|
if (mux->writing_app)
|
|
|
|
g_free (mux->writing_app);
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
G_OBJECT_CLASS (parent_class)->finalize (object);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_create_uid:
|
|
|
|
*
|
|
|
|
* Generate new unused track UID.
|
|
|
|
*
|
|
|
|
* Returns: New track UID.
|
|
|
|
*/
|
2008-08-02 17:47:32 +00:00
|
|
|
static guint64
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_matroska_mux_create_uid (void)
|
2004-09-01 12:10:21 +00:00
|
|
|
{
|
2008-08-02 17:47:32 +00:00
|
|
|
guint64 uid = 0;
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
|
2008-08-02 17:47:32 +00:00
|
|
|
G_LOCK (used_uids);
|
|
|
|
|
|
|
|
if (!used_uids)
|
|
|
|
used_uids = g_array_sized_new (FALSE, FALSE, sizeof (guint64), 10);
|
2004-09-01 12:10:21 +00:00
|
|
|
|
|
|
|
while (!uid) {
|
|
|
|
guint i;
|
|
|
|
|
2008-08-02 17:47:32 +00:00
|
|
|
uid = (((guint64) g_random_int ()) << 32) | g_random_int ();
|
2004-09-01 12:10:21 +00:00
|
|
|
for (i = 0; i < used_uids->len; i++) {
|
2008-08-02 17:47:32 +00:00
|
|
|
if (g_array_index (used_uids, guint64, i) == uid) {
|
2004-09-01 12:10:21 +00:00
|
|
|
uid = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
g_array_append_val (used_uids, uid);
|
|
|
|
}
|
2008-08-02 17:47:32 +00:00
|
|
|
|
|
|
|
G_UNLOCK (used_uids);
|
2004-09-01 12:10:21 +00:00
|
|
|
return uid;
|
|
|
|
}
|
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/**
|
2009-01-08 15:56:46 +00:00
|
|
|
* gst_matroska_pad_reset:
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
* @collect_pad: the #GstMatroskaPad
|
|
|
|
*
|
2009-01-08 15:56:46 +00:00
|
|
|
* Reset and/or release resources of a matroska collect pad.
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
*/
|
|
|
|
static void
|
2009-01-08 15:56:46 +00:00
|
|
|
gst_matroska_pad_reset (GstMatroskaPad * collect_pad, gboolean full)
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
{
|
2009-01-08 15:56:46 +00:00
|
|
|
gchar *name = NULL;
|
|
|
|
GstMatroskaTrackType type = 0;
|
2008-11-05 14:42:35 +00:00
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* free track information */
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
if (collect_pad->track != NULL) {
|
2009-01-08 15:56:46 +00:00
|
|
|
/* retrieve for optional later use */
|
|
|
|
name = collect_pad->track->name;
|
|
|
|
type = collect_pad->track->type;
|
|
|
|
/* extra for video */
|
|
|
|
if (type == GST_MATROSKA_TRACK_TYPE_VIDEO) {
|
|
|
|
GstMatroskaTrackVideoContext *ctx =
|
|
|
|
(GstMatroskaTrackVideoContext *) collect_pad->track;
|
|
|
|
|
|
|
|
if (ctx->dirac_unit) {
|
|
|
|
gst_buffer_unref (ctx->dirac_unit);
|
|
|
|
ctx->dirac_unit = NULL;
|
|
|
|
}
|
|
|
|
}
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
g_free (collect_pad->track->codec_id);
|
|
|
|
g_free (collect_pad->track->codec_name);
|
2009-01-08 15:56:46 +00:00
|
|
|
if (full)
|
|
|
|
g_free (collect_pad->track->name);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
g_free (collect_pad->track->language);
|
|
|
|
g_free (collect_pad->track->codec_priv);
|
|
|
|
g_free (collect_pad->track);
|
|
|
|
collect_pad->track = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* free cached buffer */
|
|
|
|
if (collect_pad->buffer != NULL) {
|
|
|
|
gst_buffer_unref (collect_pad->buffer);
|
|
|
|
collect_pad->buffer = NULL;
|
|
|
|
}
|
2009-01-08 15:56:46 +00:00
|
|
|
|
|
|
|
if (!full && type != 0) {
|
|
|
|
GstMatroskaTrackContext *context;
|
|
|
|
|
|
|
|
/* create a fresh context */
|
|
|
|
switch (type) {
|
|
|
|
case GST_MATROSKA_TRACK_TYPE_VIDEO:
|
|
|
|
context = (GstMatroskaTrackContext *)
|
|
|
|
g_new0 (GstMatroskaTrackVideoContext, 1);
|
|
|
|
break;
|
|
|
|
case GST_MATROSKA_TRACK_TYPE_AUDIO:
|
|
|
|
context = (GstMatroskaTrackContext *)
|
|
|
|
g_new0 (GstMatroskaTrackAudioContext, 1);
|
|
|
|
break;
|
|
|
|
case GST_MATROSKA_TRACK_TYPE_SUBTITLE:
|
|
|
|
context = (GstMatroskaTrackContext *)
|
|
|
|
g_new0 (GstMatroskaTrackSubtitleContext, 1);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
g_assert_not_reached ();
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
context->type = type;
|
|
|
|
context->name = name;
|
|
|
|
/* TODO: check default values for the context */
|
|
|
|
context->flags = GST_MATROSKA_TRACK_ENABLED | GST_MATROSKA_TRACK_DEFAULT;
|
|
|
|
collect_pad->track = context;
|
|
|
|
collect_pad->buffer = NULL;
|
|
|
|
collect_pad->duration = 0;
|
|
|
|
collect_pad->start_ts = GST_CLOCK_TIME_NONE;
|
|
|
|
collect_pad->end_ts = GST_CLOCK_TIME_NONE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_pad_free:
|
|
|
|
* @collect_pad: the #GstMatroskaPad
|
|
|
|
*
|
|
|
|
* Release resources of a matroska collect pad.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
gst_matroska_pad_free (GstMatroskaPad * collect_pad)
|
|
|
|
{
|
|
|
|
gst_matroska_pad_reset (collect_pad, TRUE);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_reset:
|
|
|
|
* @element: #GstMatroskaMux that should be reseted.
|
|
|
|
*
|
|
|
|
* Reset matroska muxer back to initial state.
|
|
|
|
*/
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_reset (GstElement * element)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstMatroskaMux *mux = GST_MATROSKA_MUX (element);
|
2005-10-14 12:43:30 +00:00
|
|
|
GSList *walk;
|
|
|
|
|
|
|
|
/* reset EBML write */
|
|
|
|
gst_ebml_write_reset (mux->ebml_write);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* reset input */
|
|
|
|
mux->state = GST_MATROSKA_MUX_STATE_START;
|
|
|
|
|
|
|
|
/* clean up existing streams */
|
2009-01-08 15:56:46 +00:00
|
|
|
|
|
|
|
for (walk = mux->collect->data; walk; walk = g_slist_next (walk)) {
|
2005-10-14 12:43:30 +00:00
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
|
|
|
|
collect_pad = (GstMatroskaPad *) walk->data;
|
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* reset collect pad to pristine state */
|
|
|
|
gst_matroska_pad_reset (collect_pad, FALSE);
|
2005-10-14 12:43:30 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* reset indexes */
|
|
|
|
mux->num_indexes = 0;
|
|
|
|
g_free (mux->index);
|
|
|
|
mux->index = NULL;
|
|
|
|
|
|
|
|
/* reset timers */
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
mux->time_scale = GST_MSECOND;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
mux->duration = 0;
|
2004-09-01 12:10:21 +00:00
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
/* reset cluster */
|
|
|
|
mux->cluster = 0;
|
|
|
|
mux->cluster_time = 0;
|
|
|
|
mux->cluster_pos = 0;
|
2004-09-09 17:54:26 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/* reset tags */
|
2009-01-06 17:48:10 +00:00
|
|
|
gst_tag_setter_reset_tags (GST_TAG_SETTER (mux));
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_handle_src_event:
|
|
|
|
* @pad: Pad which received the event.
|
|
|
|
* @event: Received event.
|
|
|
|
*
|
|
|
|
* handle events - copied from oggmux without understanding
|
|
|
|
*
|
|
|
|
* Returns: #TRUE on success.
|
|
|
|
*/
|
|
|
|
static gboolean
|
|
|
|
gst_matroska_mux_handle_src_event (GstPad * pad, GstEvent * event)
|
|
|
|
{
|
|
|
|
GstEventType type;
|
|
|
|
|
|
|
|
type = event ? GST_EVENT_TYPE (event) : GST_EVENT_UNKNOWN;
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case GST_EVENT_SEEK:
|
|
|
|
/* disable seeking for now */
|
|
|
|
return FALSE;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return gst_pad_event_default (pad, event);
|
|
|
|
}
|
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_handle_sink_event:
|
|
|
|
* @pad: Pad which received the event.
|
|
|
|
* @event: Received event.
|
|
|
|
*
|
|
|
|
* handle events - informational ones like tags
|
|
|
|
*
|
|
|
|
* Returns: #TRUE on success.
|
|
|
|
*/
|
|
|
|
static gboolean
|
|
|
|
gst_matroska_mux_handle_sink_event (GstPad * pad, GstEvent * event)
|
|
|
|
{
|
|
|
|
GstMatroskaTrackContext *context;
|
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
GstMatroskaMux *mux;
|
|
|
|
GstTagList *list;
|
2008-08-09 14:02:27 +00:00
|
|
|
gboolean ret = TRUE;
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
|
|
|
|
mux = GST_MATROSKA_MUX (gst_pad_get_parent (pad));
|
|
|
|
|
|
|
|
switch (GST_EVENT_TYPE (event)) {
|
|
|
|
case GST_EVENT_TAG:
|
2009-01-06 17:48:10 +00:00
|
|
|
GST_DEBUG_OBJECT (mux, "received tag event");
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
gst_event_parse_tag (event, &list);
|
2009-01-06 17:48:10 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
collect_pad = (GstMatroskaPad *) gst_pad_get_element_private (pad);
|
|
|
|
g_assert (collect_pad);
|
|
|
|
context = collect_pad->track;
|
|
|
|
g_assert (context);
|
|
|
|
/* FIXME ?
|
|
|
|
* strictly speaking, the incoming language code may only be 639-1, so not
|
|
|
|
* 639-2 according to matroska specs, but it will have to do for now */
|
|
|
|
gst_tag_list_get_string (list, GST_TAG_LANGUAGE_CODE, &context->language);
|
|
|
|
|
2009-01-06 17:48:10 +00:00
|
|
|
gst_tag_setter_merge_tags (GST_TAG_SETTER (mux), list,
|
|
|
|
gst_tag_setter_get_tag_merge_mode (GST_TAG_SETTER (mux)));
|
2008-08-09 14:02:27 +00:00
|
|
|
break;
|
|
|
|
case GST_EVENT_NEWSEGMENT:
|
|
|
|
/* We don't support NEWSEGMENT events */
|
|
|
|
ret = FALSE;
|
|
|
|
gst_event_unref (event);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* now GstCollectPads can take care of the rest, e.g. EOS */
|
2008-08-09 14:02:27 +00:00
|
|
|
if (ret)
|
|
|
|
ret = mux->collect_event (pad, event);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
gst_object_unref (mux);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_video_pad_setcaps:
|
|
|
|
* @pad: Pad which got the caps.
|
|
|
|
* @caps: New caps.
|
|
|
|
*
|
|
|
|
* Setcaps function for video sink pad.
|
|
|
|
*
|
|
|
|
* Returns: #TRUE on success.
|
|
|
|
*/
|
|
|
|
static gboolean
|
|
|
|
gst_matroska_mux_video_pad_setcaps (GstPad * pad, GstCaps * caps)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstMatroskaTrackContext *context = NULL;
|
|
|
|
GstMatroskaTrackVideoContext *videocontext;
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
GstMatroskaMux *mux;
|
2005-10-14 12:43:30 +00:00
|
|
|
GstMatroskaPad *collect_pad;
|
2008-01-08 19:57:23 +00:00
|
|
|
GstStructure *structure;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
const gchar *mimetype;
|
2009-01-08 15:56:46 +00:00
|
|
|
const GValue *value = NULL;
|
|
|
|
const GstBuffer *codec_buf = NULL;
|
2005-10-14 12:43:30 +00:00
|
|
|
gint width, height, pixel_width, pixel_height;
|
2008-01-08 19:57:23 +00:00
|
|
|
gint fps_d, fps_n;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
mux = GST_MATROSKA_MUX (GST_PAD_PARENT (pad));
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* find context */
|
2005-10-14 12:43:30 +00:00
|
|
|
collect_pad = (GstMatroskaPad *) gst_pad_get_element_private (pad);
|
|
|
|
g_assert (collect_pad);
|
|
|
|
context = collect_pad->track;
|
|
|
|
g_assert (context);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
g_assert (context->type == GST_MATROSKA_TRACK_TYPE_VIDEO);
|
|
|
|
videocontext = (GstMatroskaTrackVideoContext *) context;
|
|
|
|
|
|
|
|
/* gst -> matroska ID'ing */
|
2003-12-22 01:47:09 +00:00
|
|
|
structure = gst_caps_get_structure (caps, 0);
|
|
|
|
|
|
|
|
mimetype = gst_structure_get_name (structure);
|
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
if (!strcmp (mimetype, "video/x-theora")) {
|
|
|
|
/* we'll extract the details later from the theora identification header */
|
|
|
|
goto skip_details;
|
|
|
|
}
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
/* get general properties */
|
|
|
|
gst_structure_get_int (structure, "width", &width);
|
|
|
|
gst_structure_get_int (structure, "height", &height);
|
|
|
|
videocontext->pixel_width = width;
|
|
|
|
videocontext->pixel_height = height;
|
2009-02-23 11:14:23 +00:00
|
|
|
if (gst_structure_get_fraction (structure, "framerate", &fps_n, &fps_d)
|
|
|
|
&& fps_n > 0) {
|
2008-01-08 19:57:23 +00:00
|
|
|
context->default_duration =
|
|
|
|
gst_util_uint64_scale_int (GST_SECOND, fps_d, fps_n);
|
|
|
|
GST_LOG_OBJECT (pad, "default duration = %" GST_TIME_FORMAT,
|
|
|
|
GST_TIME_ARGS (context->default_duration));
|
|
|
|
} else {
|
|
|
|
context->default_duration = 0;
|
|
|
|
}
|
2005-11-28 15:13:22 +00:00
|
|
|
if (gst_structure_get_fraction (structure, "pixel-aspect-ratio",
|
|
|
|
&pixel_width, &pixel_height)) {
|
2003-12-22 01:47:09 +00:00
|
|
|
if (pixel_width > pixel_height) {
|
|
|
|
videocontext->display_width = width * pixel_width / pixel_height;
|
|
|
|
videocontext->display_height = height;
|
|
|
|
} else if (pixel_width < pixel_height) {
|
|
|
|
videocontext->display_width = width;
|
|
|
|
videocontext->display_height = height * pixel_height / pixel_width;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
} else {
|
|
|
|
videocontext->display_width = 0;
|
|
|
|
videocontext->display_height = 0;
|
|
|
|
}
|
2003-12-22 01:47:09 +00:00
|
|
|
} else {
|
|
|
|
videocontext->display_width = 0;
|
|
|
|
videocontext->display_height = 0;
|
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
skip_details:
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
videocontext->asr_mode = GST_MATROSKA_ASPECT_RATIO_MODE_FREE;
|
|
|
|
videocontext->fourcc = 0;
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: - check if we handle all codecs by the spec, i.e. codec private
|
|
|
|
* data and other settings
|
|
|
|
* - add new formats
|
|
|
|
*/
|
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* extract codec_data, may turn out needed */
|
|
|
|
value = gst_structure_get_value (structure, "codec_data");
|
|
|
|
if (value)
|
|
|
|
codec_buf = gst_value_get_buffer (value);
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
/* find type */
|
|
|
|
if (!strcmp (mimetype, "video/x-raw-yuv")) {
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_UNCOMPRESSED);
|
|
|
|
gst_structure_get_fourcc (structure, "format", &videocontext->fourcc);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2005-10-18 18:06:14 +00:00
|
|
|
} else if (!strcmp (mimetype, "image/jpeg")) {
|
2003-12-22 01:47:09 +00:00
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MJPEG);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2006-04-25 11:45:00 +00:00
|
|
|
} else if (!strcmp (mimetype, "video/x-xvid") /* MS/VfW compatibility cases */
|
|
|
|
||!strcmp (mimetype, "video/x-huffyuv")
|
|
|
|
|| !strcmp (mimetype, "video/x-divx")
|
|
|
|
|| !strcmp (mimetype, "video/x-dv")
|
2009-01-08 15:56:46 +00:00
|
|
|
|| !strcmp (mimetype, "video/x-h263")
|
2009-08-06 23:15:17 +00:00
|
|
|
|| !strcmp (mimetype, "video/x-msmpeg")
|
|
|
|
|| !strcmp (mimetype, "video/x-wmv")) {
|
2004-08-27 17:01:12 +00:00
|
|
|
BITMAPINFOHEADER *bih;
|
2006-04-25 11:45:00 +00:00
|
|
|
gint size = sizeof (BITMAPINFOHEADER);
|
2009-01-08 15:56:46 +00:00
|
|
|
guint32 fourcc = 0;
|
2003-12-22 01:47:09 +00:00
|
|
|
|
2006-04-25 11:45:00 +00:00
|
|
|
if (!strcmp (mimetype, "video/x-xvid"))
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('X', 'V', 'I', 'D');
|
2006-04-25 11:45:00 +00:00
|
|
|
else if (!strcmp (mimetype, "video/x-huffyuv"))
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('H', 'F', 'Y', 'U');
|
2006-04-25 11:45:00 +00:00
|
|
|
else if (!strcmp (mimetype, "video/x-dv"))
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('D', 'V', 'S', 'D');
|
2006-04-25 11:45:00 +00:00
|
|
|
else if (!strcmp (mimetype, "video/x-h263"))
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('H', '2', '6', '3');
|
2006-04-25 11:45:00 +00:00
|
|
|
else if (!strcmp (mimetype, "video/x-divx")) {
|
|
|
|
gint divxversion;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "divxversion", &divxversion);
|
|
|
|
switch (divxversion) {
|
|
|
|
case 3:
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('D', 'I', 'V', '3');
|
2006-04-25 11:45:00 +00:00
|
|
|
break;
|
|
|
|
case 4:
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('D', 'I', 'V', 'X');
|
2006-04-25 11:45:00 +00:00
|
|
|
break;
|
|
|
|
case 5:
|
2009-01-08 15:56:46 +00:00
|
|
|
fourcc = GST_MAKE_FOURCC ('D', 'X', '5', '0');
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
} else if (!strcmp (mimetype, "video/x-msmpeg")) {
|
|
|
|
gint msmpegversion;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "msmpegversion", &msmpegversion);
|
|
|
|
switch (msmpegversion) {
|
|
|
|
case 41:
|
|
|
|
fourcc = GST_MAKE_FOURCC ('M', 'P', 'G', '4');
|
|
|
|
break;
|
|
|
|
case 42:
|
|
|
|
fourcc = GST_MAKE_FOURCC ('M', 'P', '4', '2');
|
|
|
|
break;
|
|
|
|
case 43:
|
|
|
|
goto msmpeg43;
|
2006-04-25 11:45:00 +00:00
|
|
|
break;
|
|
|
|
}
|
2009-08-06 23:15:17 +00:00
|
|
|
} else if (!strcmp (mimetype, "video/x-wmv")) {
|
|
|
|
gint wmvversion;
|
|
|
|
guint32 format;
|
|
|
|
if (gst_structure_get_fourcc (structure, "format", &format)) {
|
|
|
|
fourcc = format;
|
|
|
|
} else if (gst_structure_get_int (structure, "wmvversion", &wmvversion)) {
|
|
|
|
if (wmvversion == 2) {
|
|
|
|
fourcc = GST_MAKE_FOURCC ('W', 'M', 'V', '2');
|
|
|
|
} else if (wmvversion == 1) {
|
|
|
|
fourcc = GST_MAKE_FOURCC ('W', 'M', 'V', '1');
|
|
|
|
} else if (wmvversion == 3) {
|
|
|
|
fourcc = GST_MAKE_FOURCC ('W', 'M', 'V', '3');
|
|
|
|
}
|
|
|
|
}
|
2003-12-22 01:47:09 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
if (!fourcc)
|
|
|
|
return FALSE;
|
2006-04-25 11:45:00 +00:00
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
bih = g_new0 (BITMAPINFOHEADER, 1);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_size, size);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_width, videocontext->pixel_width);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_height, videocontext->pixel_height);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_compression, fourcc);
|
|
|
|
GST_WRITE_UINT16_LE (&bih->bi_planes, (guint16) 1);
|
|
|
|
GST_WRITE_UINT16_LE (&bih->bi_bit_count, (guint16) 24);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_size_image, videocontext->pixel_width *
|
|
|
|
videocontext->pixel_height * 3);
|
|
|
|
|
|
|
|
/* process codec private/initialization data, if any */
|
|
|
|
if (codec_buf) {
|
|
|
|
size += GST_BUFFER_SIZE (codec_buf);
|
2006-04-25 11:45:00 +00:00
|
|
|
bih = g_realloc (bih, size);
|
|
|
|
GST_WRITE_UINT32_LE (&bih->bi_size, size);
|
|
|
|
memcpy ((guint8 *) bih + sizeof (BITMAPINFOHEADER),
|
2009-01-08 15:56:46 +00:00
|
|
|
GST_BUFFER_DATA (codec_buf), GST_BUFFER_SIZE (codec_buf));
|
2006-04-25 11:45:00 +00:00
|
|
|
}
|
2004-08-27 17:01:12 +00:00
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_VFW_FOURCC);
|
|
|
|
context->codec_priv = (gpointer) bih;
|
2006-04-25 11:45:00 +00:00
|
|
|
context->codec_priv_size = size;
|
2003-12-22 01:47:09 +00:00
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "video/x-h264")) {
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MPEG4_AVC);
|
|
|
|
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Create avcC header */
|
2009-01-08 15:56:46 +00:00
|
|
|
if (codec_buf != NULL) {
|
|
|
|
context->codec_priv_size = GST_BUFFER_SIZE (codec_buf);
|
|
|
|
context->codec_priv = g_malloc0 (context->codec_priv_size);
|
|
|
|
memcpy (context->codec_priv, GST_BUFFER_DATA (codec_buf),
|
|
|
|
context->codec_priv_size);
|
2005-10-14 12:43:30 +00:00
|
|
|
}
|
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "video/x-theora")) {
|
|
|
|
const GValue *streamheader;
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_THEORA);
|
|
|
|
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
streamheader = gst_structure_get_value (structure, "streamheader");
|
|
|
|
if (!theora_streamheader_to_codecdata (streamheader, context)) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("theora stream headers missing or malformed"));
|
|
|
|
return FALSE;
|
|
|
|
}
|
2008-02-29 06:18:55 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "video/x-dirac")) {
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_DIRAC);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "video/mpeg")) {
|
|
|
|
gint mpegversion;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "mpegversion", &mpegversion);
|
|
|
|
switch (mpegversion) {
|
|
|
|
case 1:
|
2004-03-15 19:32:27 +00:00
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MPEG1);
|
|
|
|
break;
|
2003-12-22 01:47:09 +00:00
|
|
|
case 2:
|
2004-03-15 19:32:27 +00:00
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MPEG2);
|
|
|
|
break;
|
2005-10-18 18:06:14 +00:00
|
|
|
case 4:
|
2004-03-15 19:32:27 +00:00
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MPEG4_ASP);
|
|
|
|
break;
|
2005-10-18 18:06:14 +00:00
|
|
|
default:
|
|
|
|
return FALSE;
|
2003-12-22 01:47:09 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2009-01-08 15:56:46 +00:00
|
|
|
/* global headers may be in codec data */
|
|
|
|
if (codec_buf != NULL) {
|
|
|
|
context->codec_priv_size = GST_BUFFER_SIZE (codec_buf);
|
|
|
|
context->codec_priv = g_malloc0 (context->codec_priv_size);
|
|
|
|
memcpy (context->codec_priv, GST_BUFFER_DATA (codec_buf),
|
|
|
|
context->codec_priv_size);
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "video/x-msmpeg")) {
|
2009-01-08 15:56:46 +00:00
|
|
|
msmpeg43:
|
|
|
|
/* can only make it here if preceding case verified it was version 3 */
|
2003-12-22 01:47:09 +00:00
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_MSMPEG4V3);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2008-08-25 14:15:43 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "video/x-pn-realvideo")) {
|
|
|
|
gint rmversion;
|
|
|
|
const GValue *mdpr_data;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "rmversion", &rmversion);
|
|
|
|
switch (rmversion) {
|
|
|
|
case 1:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_REALVIDEO1);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_REALVIDEO2);
|
|
|
|
break;
|
|
|
|
case 3:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_REALVIDEO3);
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_VIDEO_REALVIDEO4);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
mdpr_data = gst_structure_get_value (structure, "mdpr_data");
|
|
|
|
if (mdpr_data != NULL) {
|
|
|
|
guint8 *priv_data = NULL;
|
|
|
|
guint priv_data_size = 0;
|
|
|
|
|
|
|
|
GstBuffer *codec_data_buf = g_value_peek_pointer (mdpr_data);
|
|
|
|
|
|
|
|
priv_data_size = GST_BUFFER_SIZE (codec_data_buf);
|
|
|
|
priv_data = g_malloc0 (priv_data_size);
|
|
|
|
|
|
|
|
memcpy (priv_data, GST_BUFFER_DATA (codec_data_buf), priv_data_size);
|
|
|
|
|
|
|
|
context->codec_priv = priv_data;
|
|
|
|
context->codec_priv_size = priv_data_size;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return FALSE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2009-02-15 18:49:44 +00:00
|
|
|
/* N > 0 to expect a particular number of headers, negative if the
|
|
|
|
number of headers is variable */
|
|
|
|
static gboolean
|
|
|
|
xiphN_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context, GstBuffer ** p_buf0, int N)
|
|
|
|
{
|
|
|
|
GstBuffer **buf = NULL;
|
|
|
|
GArray *bufarr;
|
|
|
|
guint8 *priv_data;
|
|
|
|
guint bufi, i, offset, priv_data_size;
|
|
|
|
|
|
|
|
if (streamheader == NULL)
|
|
|
|
goto no_stream_headers;
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (streamheader) != GST_TYPE_ARRAY)
|
|
|
|
goto wrong_type;
|
|
|
|
|
|
|
|
bufarr = g_value_peek_pointer (streamheader);
|
|
|
|
if (bufarr->len <= 0 || bufarr->len > 255) /* at least one header, and count stored in a byte */
|
|
|
|
goto wrong_count;
|
|
|
|
if (N > 0 && bufarr->len != N)
|
|
|
|
goto wrong_count;
|
|
|
|
|
|
|
|
context->xiph_headers_to_skip = bufarr->len;
|
|
|
|
|
|
|
|
buf = (GstBuffer **) g_malloc0 (sizeof (GstBuffer *) * bufarr->len);
|
|
|
|
for (i = 0; i < bufarr->len; i++) {
|
|
|
|
GValue *bufval = &g_array_index (bufarr, GValue, i);
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER) {
|
|
|
|
g_free (buf);
|
|
|
|
goto wrong_content_type;
|
|
|
|
}
|
|
|
|
|
|
|
|
buf[i] = g_value_peek_pointer (bufval);
|
|
|
|
}
|
|
|
|
|
|
|
|
priv_data_size = 1;
|
|
|
|
if (bufarr->len > 0) {
|
|
|
|
for (i = 0; i < bufarr->len - 1; i++) {
|
|
|
|
priv_data_size += GST_BUFFER_SIZE (buf[i]) / 0xff + 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < bufarr->len; ++i) {
|
|
|
|
priv_data_size += GST_BUFFER_SIZE (buf[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
priv_data = g_malloc0 (priv_data_size);
|
|
|
|
|
|
|
|
priv_data[0] = bufarr->len - 1;
|
|
|
|
offset = 1;
|
|
|
|
|
|
|
|
if (bufarr->len > 0) {
|
|
|
|
for (bufi = 0; bufi < bufarr->len - 1; bufi++) {
|
|
|
|
for (i = 0; i < GST_BUFFER_SIZE (buf[bufi]) / 0xff; ++i) {
|
|
|
|
priv_data[offset++] = 0xff;
|
|
|
|
}
|
|
|
|
priv_data[offset++] = GST_BUFFER_SIZE (buf[bufi]) % 0xff;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < bufarr->len; ++i) {
|
|
|
|
memcpy (priv_data + offset, GST_BUFFER_DATA (buf[i]),
|
|
|
|
GST_BUFFER_SIZE (buf[i]));
|
|
|
|
offset += GST_BUFFER_SIZE (buf[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
context->codec_priv = priv_data;
|
|
|
|
context->codec_priv_size = priv_data_size;
|
|
|
|
|
|
|
|
if (p_buf0)
|
|
|
|
*p_buf0 = gst_buffer_ref (buf[0]);
|
|
|
|
|
|
|
|
g_free (buf);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
|
|
|
|
/* ERRORS */
|
|
|
|
no_stream_headers:
|
|
|
|
{
|
|
|
|
GST_WARNING ("required streamheaders missing in sink caps!");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_type:
|
|
|
|
{
|
|
|
|
GST_WARNING ("streamheaders are not a GST_TYPE_ARRAY, but a %s",
|
|
|
|
G_VALUE_TYPE_NAME (streamheader));
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_count:
|
|
|
|
{
|
|
|
|
GST_WARNING ("got %u streamheaders, not 3 as expected", bufarr->len);
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_content_type:
|
|
|
|
{
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* FIXME: after release make all code use xiph3_streamheader_to_codecdata() */
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
static gboolean
|
|
|
|
xiph3_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context, GstBuffer ** p_buf0)
|
|
|
|
{
|
|
|
|
GstBuffer *buf[3];
|
|
|
|
GArray *bufarr;
|
|
|
|
guint8 *priv_data;
|
|
|
|
guint i, offset, priv_data_size;
|
|
|
|
|
|
|
|
if (streamheader == NULL)
|
|
|
|
goto no_stream_headers;
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (streamheader) != GST_TYPE_ARRAY)
|
|
|
|
goto wrong_type;
|
|
|
|
|
|
|
|
bufarr = g_value_peek_pointer (streamheader);
|
|
|
|
if (bufarr->len != 3)
|
|
|
|
goto wrong_count;
|
|
|
|
|
|
|
|
context->xiph_headers_to_skip = bufarr->len;
|
|
|
|
|
|
|
|
for (i = 0; i < 3; i++) {
|
|
|
|
GValue *bufval = &g_array_index (bufarr, GValue, i);
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER)
|
|
|
|
goto wrong_content_type;
|
|
|
|
|
|
|
|
buf[i] = g_value_peek_pointer (bufval);
|
|
|
|
}
|
|
|
|
|
|
|
|
priv_data_size = 1;
|
|
|
|
priv_data_size += GST_BUFFER_SIZE (buf[0]) / 0xff + 1;
|
|
|
|
priv_data_size += GST_BUFFER_SIZE (buf[1]) / 0xff + 1;
|
|
|
|
|
|
|
|
for (i = 0; i < 3; ++i) {
|
|
|
|
priv_data_size += GST_BUFFER_SIZE (buf[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
priv_data = g_malloc0 (priv_data_size);
|
|
|
|
|
|
|
|
priv_data[0] = 2;
|
|
|
|
offset = 1;
|
|
|
|
|
|
|
|
for (i = 0; i < GST_BUFFER_SIZE (buf[0]) / 0xff; ++i) {
|
|
|
|
priv_data[offset++] = 0xff;
|
|
|
|
}
|
|
|
|
priv_data[offset++] = GST_BUFFER_SIZE (buf[0]) % 0xff;
|
|
|
|
|
|
|
|
for (i = 0; i < GST_BUFFER_SIZE (buf[1]) / 0xff; ++i) {
|
|
|
|
priv_data[offset++] = 0xff;
|
|
|
|
}
|
|
|
|
priv_data[offset++] = GST_BUFFER_SIZE (buf[1]) % 0xff;
|
|
|
|
|
|
|
|
for (i = 0; i < 3; ++i) {
|
|
|
|
memcpy (priv_data + offset, GST_BUFFER_DATA (buf[i]),
|
|
|
|
GST_BUFFER_SIZE (buf[i]));
|
|
|
|
offset += GST_BUFFER_SIZE (buf[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
context->codec_priv = priv_data;
|
|
|
|
context->codec_priv_size = priv_data_size;
|
|
|
|
|
|
|
|
if (p_buf0)
|
|
|
|
*p_buf0 = gst_buffer_ref (buf[0]);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
|
|
|
|
/* ERRORS */
|
|
|
|
no_stream_headers:
|
|
|
|
{
|
|
|
|
GST_WARNING ("required streamheaders missing in sink caps!");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_type:
|
|
|
|
{
|
|
|
|
GST_WARNING ("streamheaders are not a GST_TYPE_ARRAY, but a %s",
|
|
|
|
G_VALUE_TYPE_NAME (streamheader));
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_count:
|
|
|
|
{
|
|
|
|
GST_WARNING ("got %u streamheaders, not 3 as expected", bufarr->len);
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
wrong_content_type:
|
|
|
|
{
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static gboolean
|
|
|
|
vorbis_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context)
|
|
|
|
{
|
|
|
|
GstBuffer *buf0 = NULL;
|
|
|
|
|
2009-02-15 18:49:44 +00:00
|
|
|
/* FIXME: change to use xiphN_streamheader_to_codecdata() after release */
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
if (!xiph3_streamheader_to_codecdata (streamheader, context, &buf0))
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
if (buf0 == NULL || GST_BUFFER_SIZE (buf0) < 1 + 6 + 4) {
|
|
|
|
GST_WARNING ("First vorbis header too small, ignoring");
|
|
|
|
} else {
|
|
|
|
if (memcmp (GST_BUFFER_DATA (buf0) + 1, "vorbis", 6) == 0) {
|
|
|
|
GstMatroskaTrackAudioContext *audiocontext;
|
|
|
|
guint8 *hdr;
|
|
|
|
|
|
|
|
hdr = GST_BUFFER_DATA (buf0) + 1 + 6 + 4;
|
|
|
|
audiocontext = (GstMatroskaTrackAudioContext *) context;
|
|
|
|
audiocontext->channels = GST_READ_UINT8 (hdr);
|
|
|
|
audiocontext->samplerate = GST_READ_UINT32_LE (hdr + 1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (buf0)
|
|
|
|
gst_buffer_unref (buf0);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static gboolean
|
|
|
|
theora_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context)
|
|
|
|
{
|
|
|
|
GstBuffer *buf0 = NULL;
|
|
|
|
|
2009-02-15 18:49:44 +00:00
|
|
|
/* FIXME: change to use xiphN_streamheader_to_codecdata() after release */
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
if (!xiph3_streamheader_to_codecdata (streamheader, context, &buf0))
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
if (buf0 == NULL || GST_BUFFER_SIZE (buf0) < 1 + 6 + 26) {
|
|
|
|
GST_WARNING ("First theora header too small, ignoring");
|
|
|
|
} else if (memcmp (GST_BUFFER_DATA (buf0), "\200theora\003\002", 9) != 0) {
|
|
|
|
GST_WARNING ("First header not a theora identification header, ignoring");
|
|
|
|
} else {
|
|
|
|
GstMatroskaTrackVideoContext *videocontext;
|
|
|
|
guint fps_num, fps_denom, par_num, par_denom;
|
|
|
|
guint8 *hdr;
|
|
|
|
|
|
|
|
hdr = GST_BUFFER_DATA (buf0) + 1 + 6 + 3 + 2 + 2;
|
|
|
|
|
|
|
|
videocontext = (GstMatroskaTrackVideoContext *) context;
|
|
|
|
videocontext->pixel_width = GST_READ_UINT32_BE (hdr) >> 8;
|
|
|
|
videocontext->pixel_height = GST_READ_UINT32_BE (hdr + 3) >> 8;
|
|
|
|
hdr += 3 + 3 + 1 + 1;
|
|
|
|
fps_num = GST_READ_UINT32_BE (hdr);
|
|
|
|
fps_denom = GST_READ_UINT32_BE (hdr + 4);
|
|
|
|
context->default_duration = gst_util_uint64_scale_int (GST_SECOND,
|
|
|
|
fps_denom, fps_num);
|
|
|
|
hdr += 4 + 4;
|
|
|
|
par_num = GST_READ_UINT32_BE (hdr) >> 8;
|
|
|
|
par_denom = GST_READ_UINT32_BE (hdr + 3) >> 8;
|
|
|
|
if (par_num > 0 && par_num > 0) {
|
|
|
|
if (par_num > par_denom) {
|
|
|
|
videocontext->display_width =
|
|
|
|
videocontext->pixel_width * par_num / par_denom;
|
|
|
|
videocontext->display_height = videocontext->pixel_height;
|
|
|
|
} else if (par_num < par_denom) {
|
|
|
|
videocontext->display_width = videocontext->pixel_width;
|
|
|
|
videocontext->display_height =
|
|
|
|
videocontext->pixel_height * par_denom / par_num;
|
|
|
|
} else {
|
|
|
|
videocontext->display_width = 0;
|
|
|
|
videocontext->display_height = 0;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
videocontext->display_width = 0;
|
|
|
|
videocontext->display_height = 0;
|
|
|
|
}
|
|
|
|
hdr += 3 + 3;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (buf0)
|
|
|
|
gst_buffer_unref (buf0);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
2005-10-14 12:43:30 +00:00
|
|
|
|
2009-02-15 18:49:44 +00:00
|
|
|
static gboolean
|
|
|
|
kate_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context)
|
|
|
|
{
|
|
|
|
GstBuffer *buf0 = NULL;
|
|
|
|
|
|
|
|
if (!xiphN_streamheader_to_codecdata (streamheader, context, &buf0, -1))
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
if (buf0 == NULL || GST_BUFFER_SIZE (buf0) < 64) { /* Kate ID header is 64 bytes */
|
|
|
|
GST_WARNING ("First kate header too small, ignoring");
|
|
|
|
} else if (memcmp (GST_BUFFER_DATA (buf0), "\200kate\0\0\0", 8) != 0) {
|
|
|
|
GST_WARNING ("First header not a kate identification header, ignoring");
|
|
|
|
}
|
|
|
|
|
|
|
|
if (buf0)
|
|
|
|
gst_buffer_unref (buf0);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2008-08-09 13:48:22 +00:00
|
|
|
static gboolean
|
|
|
|
flac_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context)
|
|
|
|
{
|
|
|
|
GArray *bufarr;
|
|
|
|
gint i;
|
|
|
|
GValue *bufval;
|
|
|
|
GstBuffer *buffer;
|
|
|
|
|
|
|
|
if (streamheader == NULL || G_VALUE_TYPE (streamheader) != GST_TYPE_ARRAY) {
|
|
|
|
GST_WARNING ("No or invalid streamheader field in the caps");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
bufarr = g_value_peek_pointer (streamheader);
|
|
|
|
if (bufarr->len < 2) {
|
|
|
|
GST_WARNING ("Too few headers in streamheader field");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
2008-08-09 14:02:27 +00:00
|
|
|
context->xiph_headers_to_skip = bufarr->len + 1;
|
2008-08-09 13:48:22 +00:00
|
|
|
|
|
|
|
bufval = &g_array_index (bufarr, GValue, 0);
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER) {
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
buffer = g_value_peek_pointer (bufval);
|
|
|
|
|
|
|
|
/* Need at least OggFLAC mapping header, fLaC marker and STREAMINFO block */
|
|
|
|
if (GST_BUFFER_SIZE (buffer) < 9 + 4 + 4 + 34
|
|
|
|
|| memcmp (GST_BUFFER_DATA (buffer) + 1, "FLAC", 4) != 0
|
|
|
|
|| memcmp (GST_BUFFER_DATA (buffer) + 9, "fLaC", 4) != 0) {
|
|
|
|
GST_WARNING ("Invalid streamheader for FLAC");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
context->codec_priv = g_malloc (GST_BUFFER_SIZE (buffer) - 9);
|
|
|
|
context->codec_priv_size = GST_BUFFER_SIZE (buffer) - 9;
|
|
|
|
memcpy (context->codec_priv, GST_BUFFER_DATA (buffer) + 9,
|
|
|
|
GST_BUFFER_SIZE (buffer) - 9);
|
|
|
|
|
|
|
|
for (i = 1; i < bufarr->len; i++) {
|
|
|
|
bufval = &g_array_index (bufarr, GValue, i);
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
buffer = g_value_peek_pointer (bufval);
|
|
|
|
|
|
|
|
context->codec_priv =
|
|
|
|
g_realloc (context->codec_priv,
|
|
|
|
context->codec_priv_size + GST_BUFFER_SIZE (buffer));
|
2009-01-30 17:26:19 +00:00
|
|
|
memcpy ((guint8 *) context->codec_priv + context->codec_priv_size,
|
2008-08-09 13:48:22 +00:00
|
|
|
GST_BUFFER_DATA (buffer), GST_BUFFER_SIZE (buffer));
|
|
|
|
context->codec_priv_size =
|
|
|
|
context->codec_priv_size + GST_BUFFER_SIZE (buffer);
|
|
|
|
}
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2009-04-13 12:03:03 +00:00
|
|
|
static gboolean
|
|
|
|
speex_streamheader_to_codecdata (const GValue * streamheader,
|
|
|
|
GstMatroskaTrackContext * context)
|
|
|
|
{
|
|
|
|
GArray *bufarr;
|
|
|
|
GValue *bufval;
|
|
|
|
GstBuffer *buffer;
|
|
|
|
|
|
|
|
if (streamheader == NULL || G_VALUE_TYPE (streamheader) != GST_TYPE_ARRAY) {
|
|
|
|
GST_WARNING ("No or invalid streamheader field in the caps");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
bufarr = g_value_peek_pointer (streamheader);
|
|
|
|
if (bufarr->len != 2) {
|
|
|
|
GST_WARNING ("Too few headers in streamheader field");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
context->xiph_headers_to_skip = bufarr->len + 1;
|
|
|
|
|
|
|
|
bufval = &g_array_index (bufarr, GValue, 0);
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER) {
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
buffer = g_value_peek_pointer (bufval);
|
|
|
|
|
|
|
|
if (GST_BUFFER_SIZE (buffer) < 80
|
|
|
|
|| memcmp (GST_BUFFER_DATA (buffer), "Speex ", 8) != 0) {
|
|
|
|
GST_WARNING ("Invalid streamheader for Speex");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
context->codec_priv = g_malloc (GST_BUFFER_SIZE (buffer));
|
|
|
|
context->codec_priv_size = GST_BUFFER_SIZE (buffer);
|
|
|
|
memcpy (context->codec_priv, GST_BUFFER_DATA (buffer),
|
|
|
|
GST_BUFFER_SIZE (buffer));
|
|
|
|
|
|
|
|
bufval = &g_array_index (bufarr, GValue, 1);
|
|
|
|
|
|
|
|
if (G_VALUE_TYPE (bufval) != GST_TYPE_BUFFER) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
GST_WARNING ("streamheaders array does not contain GstBuffers");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
buffer = g_value_peek_pointer (bufval);
|
|
|
|
|
|
|
|
context->codec_priv =
|
|
|
|
g_realloc (context->codec_priv,
|
|
|
|
context->codec_priv_size + GST_BUFFER_SIZE (buffer));
|
|
|
|
memcpy ((guint8 *) context->codec_priv + context->codec_priv_size,
|
|
|
|
GST_BUFFER_DATA (buffer), GST_BUFFER_SIZE (buffer));
|
|
|
|
context->codec_priv_size =
|
|
|
|
context->codec_priv_size + GST_BUFFER_SIZE (buffer);
|
|
|
|
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2008-11-10 16:44:45 +00:00
|
|
|
static gchar *
|
|
|
|
aac_codec_data_to_codec_id (const GstBuffer * buf)
|
|
|
|
{
|
|
|
|
gchar *result;
|
|
|
|
gint profile;
|
|
|
|
|
|
|
|
/* default to MAIN */
|
|
|
|
profile = 1;
|
|
|
|
|
|
|
|
if (GST_BUFFER_SIZE (buf) >= 2) {
|
|
|
|
profile = GST_READ_UINT8 (GST_BUFFER_DATA (buf));
|
|
|
|
profile >>= 3;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (profile) {
|
|
|
|
case 1:
|
|
|
|
result = "MAIN";
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
result = "LC";
|
|
|
|
break;
|
|
|
|
case 3:
|
|
|
|
result = "SSR";
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
result = "LTP";
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
GST_WARNING ("unknown AAC profile, defaulting to MAIN");
|
|
|
|
result = "MAIN";
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_audio_pad_setcaps:
|
|
|
|
* @pad: Pad which got the caps.
|
|
|
|
* @caps: New caps.
|
|
|
|
*
|
|
|
|
* Setcaps function for audio sink pad.
|
|
|
|
*
|
|
|
|
* Returns: #TRUE on success.
|
|
|
|
*/
|
|
|
|
static gboolean
|
|
|
|
gst_matroska_mux_audio_pad_setcaps (GstPad * pad, GstCaps * caps)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstMatroskaTrackContext *context = NULL;
|
|
|
|
GstMatroskaTrackAudioContext *audiocontext;
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
GstMatroskaMux *mux;
|
2005-10-14 12:43:30 +00:00
|
|
|
GstMatroskaPad *collect_pad;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
const gchar *mimetype;
|
2005-10-20 19:14:27 +00:00
|
|
|
gint samplerate = 0, channels = 0;
|
2003-12-22 01:47:09 +00:00
|
|
|
GstStructure *structure;
|
2009-08-08 03:47:48 +00:00
|
|
|
const GValue *codec_data = NULL;
|
|
|
|
const GstBuffer *buf = NULL;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
mux = GST_MATROSKA_MUX (GST_PAD_PARENT (pad));
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* find context */
|
2005-10-14 12:43:30 +00:00
|
|
|
collect_pad = (GstMatroskaPad *) gst_pad_get_element_private (pad);
|
|
|
|
g_assert (collect_pad);
|
|
|
|
context = collect_pad->track;
|
|
|
|
g_assert (context);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
g_assert (context->type == GST_MATROSKA_TRACK_TYPE_AUDIO);
|
|
|
|
audiocontext = (GstMatroskaTrackAudioContext *) context;
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
structure = gst_caps_get_structure (caps, 0);
|
|
|
|
mimetype = gst_structure_get_name (structure);
|
|
|
|
|
|
|
|
/* general setup */
|
|
|
|
gst_structure_get_int (structure, "rate", &samplerate);
|
|
|
|
gst_structure_get_int (structure, "channels", &channels);
|
|
|
|
|
|
|
|
audiocontext->samplerate = samplerate;
|
|
|
|
audiocontext->channels = channels;
|
|
|
|
audiocontext->bitdepth = 0;
|
2004-09-01 12:10:21 +00:00
|
|
|
context->default_duration = 0;
|
2003-12-22 01:47:09 +00:00
|
|
|
|
2009-08-08 03:47:48 +00:00
|
|
|
codec_data = gst_structure_get_value (structure, "codec_data");
|
|
|
|
if (codec_data)
|
|
|
|
buf = gst_value_get_buffer (codec_data);
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: - check if we handle all codecs by the spec, i.e. codec private
|
|
|
|
* data and other settings
|
|
|
|
* - add new formats
|
|
|
|
*/
|
|
|
|
|
2003-12-22 01:47:09 +00:00
|
|
|
if (!strcmp (mimetype, "audio/mpeg")) {
|
|
|
|
gint mpegversion = 0;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "mpegversion", &mpegversion);
|
|
|
|
switch (mpegversion) {
|
2004-03-14 22:34:33 +00:00
|
|
|
case 1:{
|
2004-03-15 19:32:27 +00:00
|
|
|
gint layer;
|
2008-08-02 18:11:32 +00:00
|
|
|
gint version = 1;
|
|
|
|
gint spf;
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_structure_get_int (structure, "layer", &layer);
|
2008-08-02 18:11:32 +00:00
|
|
|
|
|
|
|
if (!gst_structure_get_int (structure, "mpegaudioversion", &version)) {
|
|
|
|
GST_WARNING_OBJECT (mux,
|
|
|
|
"Unable to determine MPEG audio version, assuming 1");
|
|
|
|
version = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (layer == 1)
|
|
|
|
spf = 384;
|
|
|
|
else if (layer == 2)
|
|
|
|
spf = 1152;
|
|
|
|
else if (version == 2)
|
|
|
|
spf = 576;
|
|
|
|
else
|
|
|
|
spf = 1152;
|
|
|
|
|
|
|
|
context->default_duration =
|
|
|
|
gst_util_uint64_scale (GST_SECOND, spf, audiocontext->samplerate);
|
|
|
|
|
2004-03-15 19:32:27 +00:00
|
|
|
switch (layer) {
|
|
|
|
case 1:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_MPEG1_L1);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_MPEG1_L2);
|
|
|
|
break;
|
|
|
|
case 3:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_MPEG1_L3);
|
|
|
|
break;
|
2005-10-18 18:06:14 +00:00
|
|
|
default:
|
|
|
|
return FALSE;
|
2004-03-15 19:32:27 +00:00
|
|
|
}
|
|
|
|
break;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
2003-12-22 01:47:09 +00:00
|
|
|
case 2:
|
2008-11-10 16:44:45 +00:00
|
|
|
if (buf) {
|
|
|
|
context->codec_id =
|
|
|
|
g_strdup_printf (GST_MATROSKA_CODEC_ID_AUDIO_AAC_MPEG2 "%s",
|
|
|
|
aac_codec_data_to_codec_id (buf));
|
|
|
|
} else {
|
|
|
|
GST_DEBUG_OBJECT (mux, "no AAC codec_data; not packetized");
|
|
|
|
return FALSE;
|
|
|
|
}
|
2004-03-15 19:32:27 +00:00
|
|
|
break;
|
2003-12-22 01:47:09 +00:00
|
|
|
case 4:
|
2008-11-10 16:44:45 +00:00
|
|
|
if (buf) {
|
|
|
|
context->codec_id =
|
|
|
|
g_strdup_printf (GST_MATROSKA_CODEC_ID_AUDIO_AAC_MPEG4 "%s",
|
|
|
|
aac_codec_data_to_codec_id (buf));
|
|
|
|
} else {
|
|
|
|
GST_DEBUG_OBJECT (mux, "no AAC codec_data; not packetized");
|
|
|
|
return FALSE;
|
|
|
|
}
|
2004-03-15 19:32:27 +00:00
|
|
|
break;
|
2005-10-18 18:06:14 +00:00
|
|
|
default:
|
|
|
|
return FALSE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
2003-12-22 01:47:09 +00:00
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-raw-int")) {
|
2009-03-31 13:25:58 +00:00
|
|
|
gint width, depth;
|
|
|
|
gint endianness = G_LITTLE_ENDIAN;
|
|
|
|
gboolean signedness = TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
|
2006-12-24 11:24:59 +00:00
|
|
|
if (!gst_structure_get_int (structure, "width", &width) ||
|
2005-10-20 19:14:27 +00:00
|
|
|
!gst_structure_get_int (structure, "depth", &depth) ||
|
2006-12-24 11:24:59 +00:00
|
|
|
!gst_structure_get_boolean (structure, "signed", &signedness)) {
|
|
|
|
GST_DEBUG_OBJECT (mux, "broken caps, width/depth/signed field missing");
|
2005-10-20 19:14:27 +00:00
|
|
|
return FALSE;
|
2006-12-24 11:24:59 +00:00
|
|
|
}
|
2005-10-20 19:14:27 +00:00
|
|
|
|
2006-12-24 11:24:59 +00:00
|
|
|
if (depth > 8 &&
|
|
|
|
!gst_structure_get_int (structure, "endianness", &endianness)) {
|
|
|
|
GST_DEBUG_OBJECT (mux, "broken caps, no endianness specified");
|
2005-10-14 12:43:30 +00:00
|
|
|
return FALSE;
|
2006-12-24 11:24:59 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if (width != depth) {
|
|
|
|
GST_DEBUG_OBJECT (mux, "width must be same as depth!");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
2008-08-02 17:52:16 +00:00
|
|
|
/* FIXME: where is this spec'ed out? (tpm) */
|
|
|
|
if ((width == 8 && signedness) || (width >= 16 && !signedness)) {
|
2006-12-24 11:24:59 +00:00
|
|
|
GST_DEBUG_OBJECT (mux, "8-bit PCM must be unsigned, 16-bit PCM signed");
|
|
|
|
return FALSE;
|
|
|
|
}
|
2003-12-22 01:47:09 +00:00
|
|
|
|
|
|
|
audiocontext->bitdepth = depth;
|
|
|
|
if (endianness == G_BIG_ENDIAN)
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_PCM_INT_BE);
|
|
|
|
else
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_PCM_INT_LE);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-raw-float")) {
|
2008-08-02 17:52:16 +00:00
|
|
|
gint width;
|
|
|
|
|
|
|
|
if (!gst_structure_get_int (structure, "width", &width)) {
|
|
|
|
GST_DEBUG_OBJECT (mux, "broken caps, width field missing");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
audiocontext->bitdepth = width;
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_PCM_FLOAT);
|
|
|
|
|
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-vorbis")) {
|
2005-10-14 12:43:30 +00:00
|
|
|
const GValue *streamheader;
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_VORBIS);
|
|
|
|
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
streamheader = gst_structure_get_value (structure, "streamheader");
|
|
|
|
if (!vorbis_streamheader_to_codecdata (streamheader, context)) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("vorbis stream headers missing or malformed"));
|
|
|
|
return FALSE;
|
2005-10-14 12:43:30 +00:00
|
|
|
}
|
|
|
|
return TRUE;
|
2008-08-09 13:48:22 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-flac")) {
|
|
|
|
const GValue *streamheader;
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_FLAC);
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
streamheader = gst_structure_get_value (structure, "streamheader");
|
|
|
|
if (!flac_streamheader_to_codecdata (streamheader, context)) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("flac stream headers missing or malformed"));
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
return TRUE;
|
2009-04-13 12:03:03 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-speex")) {
|
|
|
|
const GValue *streamheader;
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_SPEEX);
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
streamheader = gst_structure_get_value (structure, "streamheader");
|
|
|
|
if (!speex_streamheader_to_codecdata (streamheader, context)) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("speex stream headers missing or malformed"));
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
return TRUE;
|
2003-12-22 01:47:09 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-ac3")) {
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_AC3);
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
2004-09-15 19:43:37 +00:00
|
|
|
} else if (!strcmp (mimetype, "audio/x-tta")) {
|
2004-08-26 11:10:16 +00:00
|
|
|
gint width;
|
|
|
|
|
2004-09-01 12:10:21 +00:00
|
|
|
/* TTA frame duration */
|
|
|
|
context->default_duration = 1.04489795918367346939 * GST_SECOND;
|
|
|
|
|
2004-08-26 11:10:16 +00:00
|
|
|
gst_structure_get_int (structure, "width", &width);
|
|
|
|
audiocontext->bitdepth = width;
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_TTA);
|
|
|
|
|
2008-08-25 14:15:43 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "audio/x-pn-realaudio")) {
|
|
|
|
gint raversion;
|
|
|
|
const GValue *mdpr_data;
|
|
|
|
|
|
|
|
gst_structure_get_int (structure, "raversion", &raversion);
|
|
|
|
switch (raversion) {
|
|
|
|
case 1:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_REAL_14_4);
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_REAL_28_8);
|
|
|
|
break;
|
|
|
|
case 8:
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_REAL_COOK);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
mdpr_data = gst_structure_get_value (structure, "mdpr_data");
|
|
|
|
if (mdpr_data != NULL) {
|
|
|
|
guint8 *priv_data = NULL;
|
|
|
|
guint priv_data_size = 0;
|
|
|
|
|
|
|
|
GstBuffer *codec_data_buf = g_value_peek_pointer (mdpr_data);
|
|
|
|
|
|
|
|
priv_data_size = GST_BUFFER_SIZE (codec_data_buf);
|
|
|
|
priv_data = g_malloc0 (priv_data_size);
|
|
|
|
|
|
|
|
memcpy (priv_data, GST_BUFFER_DATA (codec_data_buf), priv_data_size);
|
|
|
|
|
|
|
|
context->codec_priv = priv_data;
|
|
|
|
context->codec_priv_size = priv_data_size;
|
|
|
|
}
|
|
|
|
|
2009-08-08 03:47:48 +00:00
|
|
|
return TRUE;
|
|
|
|
} else if (!strcmp (mimetype, "audio/x-wma")) {
|
|
|
|
guint8 *codec_priv;
|
|
|
|
guint codec_priv_size;
|
|
|
|
guint16 format;
|
|
|
|
gint block_align;
|
|
|
|
gint bitrate;
|
|
|
|
gint wmaversion;
|
|
|
|
gint depth;
|
|
|
|
|
|
|
|
if (!gst_structure_get_int (structure, "wmaversion", &wmaversion)
|
|
|
|
|| !gst_structure_get_int (structure, "block_align", &block_align)
|
|
|
|
|| !gst_structure_get_int (structure, "bitrate", &bitrate)
|
|
|
|
|| samplerate == 0 || channels == 0) {
|
|
|
|
GST_WARNING_OBJECT (mux, "Missing wmaversion/block_align/bitrate/"
|
|
|
|
"channels/rate on WMA caps");
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (wmaversion) {
|
|
|
|
case 1:
|
|
|
|
format = GST_RIFF_WAVE_FORMAT_WMAV1;
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
format = GST_RIFF_WAVE_FORMAT_WMAV2;
|
|
|
|
break;
|
|
|
|
case 3:
|
|
|
|
format = GST_RIFF_WAVE_FORMAT_WMAV3;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
GST_WARNING_OBJECT (mux, "Unexpected WMA version: %d", wmaversion);
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (gst_structure_get_int (structure, "depth", &depth))
|
|
|
|
audiocontext->bitdepth = depth;
|
|
|
|
|
|
|
|
codec_priv_size = WAVEFORMATEX_SIZE;
|
|
|
|
if (buf)
|
|
|
|
codec_priv_size += GST_BUFFER_SIZE (buf);
|
|
|
|
|
|
|
|
/* serialize waveformatex structure */
|
|
|
|
codec_priv = g_malloc0 (codec_priv_size);
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv, format);
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv + 2, channels);
|
|
|
|
GST_WRITE_UINT32_LE (codec_priv + 4, samplerate);
|
|
|
|
GST_WRITE_UINT32_LE (codec_priv + 8, bitrate / 8);
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv + 12, block_align);
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv + 14, 0);
|
|
|
|
if (buf)
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv + 16, GST_BUFFER_SIZE (buf));
|
|
|
|
else
|
|
|
|
GST_WRITE_UINT16_LE (codec_priv + 16, 0);
|
|
|
|
|
|
|
|
/* process codec private/initialization data, if any */
|
|
|
|
if (buf) {
|
|
|
|
memcpy ((guint8 *) codec_priv + WAVEFORMATEX_SIZE,
|
|
|
|
GST_BUFFER_DATA (buf), GST_BUFFER_SIZE (buf));
|
|
|
|
}
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_AUDIO_ACM);
|
|
|
|
context->codec_priv = (gpointer) codec_priv;
|
|
|
|
context->codec_priv_size = codec_priv_size;
|
2005-10-14 12:43:30 +00:00
|
|
|
return TRUE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return FALSE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_subtitle_pad_setcaps:
|
|
|
|
* @pad: Pad which got the caps.
|
|
|
|
* @caps: New caps.
|
|
|
|
*
|
|
|
|
* Setcaps function for subtitle sink pad.
|
|
|
|
*
|
|
|
|
* Returns: #TRUE on success.
|
|
|
|
*/
|
|
|
|
static gboolean
|
|
|
|
gst_matroska_mux_subtitle_pad_setcaps (GstPad * pad, GstCaps * caps)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* FIXME:
|
|
|
|
* Consider this as boilerplate code for now. There is
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
* no single subtitle creation element in GStreamer,
|
|
|
|
* neither do I know how subtitling works at all. */
|
|
|
|
|
2009-02-15 18:49:44 +00:00
|
|
|
/* There is now (at least) one such alement (kateenc), and I'm going
|
|
|
|
to handle it here and claim it works when it can be piped back
|
|
|
|
through GStreamer and VLC */
|
|
|
|
|
|
|
|
GstMatroskaTrackContext *context = NULL;
|
|
|
|
GstMatroskaTrackSubtitleContext *scontext;
|
|
|
|
GstMatroskaMux *mux;
|
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
const gchar *mimetype;
|
|
|
|
GstStructure *structure;
|
|
|
|
|
|
|
|
mux = GST_MATROSKA_MUX (GST_PAD_PARENT (pad));
|
|
|
|
|
|
|
|
/* find context */
|
|
|
|
collect_pad = (GstMatroskaPad *) gst_pad_get_element_private (pad);
|
|
|
|
g_assert (collect_pad);
|
|
|
|
context = collect_pad->track;
|
|
|
|
g_assert (context);
|
|
|
|
g_assert (context->type == GST_MATROSKA_TRACK_TYPE_SUBTITLE);
|
|
|
|
scontext = (GstMatroskaTrackSubtitleContext *) context;
|
|
|
|
|
|
|
|
structure = gst_caps_get_structure (caps, 0);
|
|
|
|
mimetype = gst_structure_get_name (structure);
|
|
|
|
|
|
|
|
/* general setup */
|
|
|
|
scontext->check_utf8 = 1;
|
|
|
|
scontext->invalid_utf8 = 0;
|
|
|
|
context->default_duration = 0;
|
|
|
|
|
|
|
|
/* TODO: - other format than Kate */
|
|
|
|
|
|
|
|
if (!strcmp (mimetype, "subtitle/x-kate")) {
|
|
|
|
const GValue *streamheader;
|
|
|
|
|
|
|
|
context->codec_id = g_strdup (GST_MATROSKA_CODEC_ID_SUBTITLE_KATE);
|
|
|
|
|
|
|
|
if (context->codec_priv != NULL) {
|
|
|
|
g_free (context->codec_priv);
|
|
|
|
context->codec_priv = NULL;
|
|
|
|
context->codec_priv_size = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
streamheader = gst_structure_get_value (structure, "streamheader");
|
|
|
|
if (!kate_streamheader_to_codecdata (streamheader, context)) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("kate stream headers missing or malformed"));
|
|
|
|
return FALSE;
|
|
|
|
}
|
|
|
|
return TRUE;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return FALSE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_request_new_pad:
|
|
|
|
* @element: #GstMatroskaMux.
|
|
|
|
* @templ: #GstPadTemplate.
|
|
|
|
* @pad_name: New pad name.
|
|
|
|
*
|
|
|
|
* Request pad function for sink templates.
|
|
|
|
*
|
|
|
|
* Returns: New #GstPad.
|
|
|
|
*/
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static GstPad *
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_request_new_pad (GstElement * element,
|
|
|
|
GstPadTemplate * templ, const gchar * pad_name)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstElementClass *klass = GST_ELEMENT_GET_CLASS (element);
|
|
|
|
GstMatroskaMux *mux = GST_MATROSKA_MUX (element);
|
2005-10-14 12:43:30 +00:00
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
GstPad *newpad = NULL;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gchar *name = NULL;
|
2005-10-14 12:43:30 +00:00
|
|
|
GstPadSetCapsFunction setcapsfunc = NULL;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstMatroskaTrackContext *context = NULL;
|
|
|
|
|
|
|
|
if (templ == gst_element_class_get_pad_template (klass, "audio_%d")) {
|
|
|
|
name = g_strdup_printf ("audio_%d", mux->num_a_streams++);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
setcapsfunc = GST_DEBUG_FUNCPTR (gst_matroska_mux_audio_pad_setcaps);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context = (GstMatroskaTrackContext *)
|
2004-03-15 19:32:27 +00:00
|
|
|
g_new0 (GstMatroskaTrackAudioContext, 1);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context->type = GST_MATROSKA_TRACK_TYPE_AUDIO;
|
|
|
|
context->name = g_strdup ("Audio");
|
|
|
|
} else if (templ == gst_element_class_get_pad_template (klass, "video_%d")) {
|
|
|
|
name = g_strdup_printf ("video_%d", mux->num_v_streams++);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
setcapsfunc = GST_DEBUG_FUNCPTR (gst_matroska_mux_video_pad_setcaps);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context = (GstMatroskaTrackContext *)
|
2004-03-15 19:32:27 +00:00
|
|
|
g_new0 (GstMatroskaTrackVideoContext, 1);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context->type = GST_MATROSKA_TRACK_TYPE_VIDEO;
|
|
|
|
context->name = g_strdup ("Video");
|
|
|
|
} else if (templ == gst_element_class_get_pad_template (klass, "subtitle_%d")) {
|
|
|
|
name = g_strdup_printf ("subtitle_%d", mux->num_t_streams++);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
setcapsfunc = GST_DEBUG_FUNCPTR (gst_matroska_mux_subtitle_pad_setcaps);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context = (GstMatroskaTrackContext *)
|
2004-03-15 19:32:27 +00:00
|
|
|
g_new0 (GstMatroskaTrackSubtitleContext, 1);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
context->type = GST_MATROSKA_TRACK_TYPE_SUBTITLE;
|
|
|
|
context->name = g_strdup ("Subtitle");
|
|
|
|
} else {
|
2005-10-14 12:43:30 +00:00
|
|
|
GST_WARNING_OBJECT (mux, "This is not our template!");
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
newpad = gst_pad_new_from_template (templ, name);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
g_free (name);
|
2005-10-14 12:43:30 +00:00
|
|
|
collect_pad = (GstMatroskaPad *)
|
2008-10-27 08:45:11 +00:00
|
|
|
gst_collect_pads_add_pad_full (mux->collect, newpad,
|
|
|
|
sizeof (GstMatroskaPad),
|
|
|
|
(GstCollectDataDestroyNotify) gst_matroska_pad_free);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
collect_pad->track = context;
|
2009-01-08 15:56:46 +00:00
|
|
|
gst_matroska_pad_reset (collect_pad, FALSE);
|
2004-03-14 22:34:33 +00:00
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/* FIXME: hacked way to override/extend the event function of
|
|
|
|
* GstCollectPads; because it sets its own event function giving the
|
|
|
|
* element no access to events.
|
|
|
|
* TODO GstCollectPads should really give its 'users' a clean chance to
|
|
|
|
* properly handle events that are not meant for collectpads itself.
|
|
|
|
* Perhaps a callback or so, though rejected (?) in #340060.
|
|
|
|
* This would allow (clean) transcoding of info from demuxer/streams
|
|
|
|
* to another muxer */
|
|
|
|
mux->collect_event = (GstPadEventFunction) GST_PAD_EVENTFUNC (newpad);
|
|
|
|
gst_pad_set_event_function (newpad,
|
|
|
|
GST_DEBUG_FUNCPTR (gst_matroska_mux_handle_sink_event));
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_pad_set_setcaps_function (newpad, setcapsfunc);
|
2006-11-16 15:36:48 +00:00
|
|
|
gst_pad_set_active (newpad, TRUE);
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_element_add_pad (element, newpad);
|
2008-06-11 11:27:46 +00:00
|
|
|
mux->num_streams++;
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
return newpad;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_release_pad:
|
|
|
|
* @element: #GstMatroskaMux.
|
|
|
|
* @pad: Pad to release.
|
|
|
|
*
|
|
|
|
* Release a previously requested pad.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
gst_matroska_mux_release_pad (GstElement * element, GstPad * pad)
|
|
|
|
{
|
|
|
|
GstMatroskaMux *mux;
|
|
|
|
GSList *walk;
|
|
|
|
|
|
|
|
mux = GST_MATROSKA_MUX (GST_PAD_PARENT (pad));
|
|
|
|
|
|
|
|
for (walk = mux->collect->data; walk; walk = g_slist_next (walk)) {
|
|
|
|
GstCollectData *cdata = (GstCollectData *) walk->data;
|
|
|
|
GstMatroskaPad *collect_pad = (GstMatroskaPad *) cdata;
|
|
|
|
|
|
|
|
if (cdata->pad == pad) {
|
2008-01-08 19:57:23 +00:00
|
|
|
GstClockTime min_dur; /* observed minimum duration */
|
|
|
|
|
2008-03-03 22:01:56 +00:00
|
|
|
if (GST_CLOCK_TIME_IS_VALID (collect_pad->start_ts) &&
|
|
|
|
GST_CLOCK_TIME_IS_VALID (collect_pad->end_ts)) {
|
|
|
|
min_dur = GST_CLOCK_DIFF (collect_pad->start_ts, collect_pad->end_ts);
|
|
|
|
if (collect_pad->duration < min_dur)
|
|
|
|
collect_pad->duration = min_dur;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (GST_CLOCK_TIME_IS_VALID (collect_pad->duration) &&
|
|
|
|
mux->duration < collect_pad->duration)
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
mux->duration = collect_pad->duration;
|
2008-03-03 22:01:56 +00:00
|
|
|
|
2008-10-27 08:45:11 +00:00
|
|
|
break;
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-10-27 08:45:11 +00:00
|
|
|
gst_collect_pads_remove_pad (mux->collect, pad);
|
|
|
|
if (gst_element_remove_pad (element, pad))
|
|
|
|
mux->num_streams--;
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_track_header:
|
|
|
|
* @mux: #GstMatroskaMux
|
|
|
|
* @context: Tack context.
|
|
|
|
*
|
|
|
|
* Write a track header.
|
|
|
|
*/
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_track_header (GstMatroskaMux * mux,
|
|
|
|
GstMatroskaTrackContext * context)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstEbmlWrite *ebml = mux->ebml_write;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
guint64 master;
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: check if everything necessary is written and check default values */
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* track type goes before the type-specific stuff */
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_TRACKNUMBER, context->num);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_TRACKTYPE, context->type);
|
|
|
|
|
2004-09-01 12:10:21 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_TRACKUID,
|
|
|
|
gst_matroska_mux_create_uid ());
|
|
|
|
if (context->default_duration) {
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_TRACKDEFAULTDURATION,
|
|
|
|
context->default_duration);
|
|
|
|
}
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
if (context->language) {
|
|
|
|
gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_TRACKLANGUAGE,
|
|
|
|
context->language);
|
|
|
|
}
|
2004-09-01 12:10:21 +00:00
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* type-specific stuff */
|
|
|
|
switch (context->type) {
|
2004-03-14 22:34:33 +00:00
|
|
|
case GST_MATROSKA_TRACK_TYPE_VIDEO:{
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstMatroskaTrackVideoContext *videocontext =
|
2004-03-15 19:32:27 +00:00
|
|
|
(GstMatroskaTrackVideoContext *) context;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TRACKVIDEO);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_VIDEOPIXELWIDTH,
|
2004-03-15 19:32:27 +00:00
|
|
|
videocontext->pixel_width);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_VIDEOPIXELHEIGHT,
|
2004-03-15 19:32:27 +00:00
|
|
|
videocontext->pixel_height);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (videocontext->display_width && videocontext->display_height) {
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_VIDEODISPLAYWIDTH,
|
|
|
|
videocontext->display_width);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_VIDEODISPLAYHEIGHT,
|
|
|
|
videocontext->display_height);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
if (context->flags & GST_MATROSKA_VIDEOTRACK_INTERLACED)
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_VIDEOFLAGINTERLACED, 1);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (videocontext->fourcc) {
|
2004-03-15 19:32:27 +00:00
|
|
|
guint32 fcc_le = GUINT32_TO_LE (videocontext->fourcc);
|
2004-03-14 22:34:33 +00:00
|
|
|
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_binary (ebml, GST_MATROSKA_ID_VIDEOCOLOURSPACE,
|
|
|
|
(gpointer) & fcc_le, 4);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
case GST_MATROSKA_TRACK_TYPE_AUDIO:{
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstMatroskaTrackAudioContext *audiocontext =
|
2004-03-15 19:32:27 +00:00
|
|
|
(GstMatroskaTrackAudioContext *) context;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TRACKAUDIO);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (audiocontext->samplerate != 8000)
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_float (ebml, GST_MATROSKA_ID_AUDIOSAMPLINGFREQ,
|
|
|
|
audiocontext->samplerate);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (audiocontext->channels != 1)
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_AUDIOCHANNELS,
|
|
|
|
audiocontext->channels);
|
configure.ac: Improve mpeg2enc detection. This is for distributions that do ship mjpegtools, but without mpeg2enc. Al...
Original commit message from CVS:
2003-12-21 Ronald Bultje <rbultje@ronald.bitfreak.net>
* configure.ac:
Improve mpeg2enc detection. This is for distributions that do
ship mjpegtools, but without mpeg2enc. Also does object check
for might there ever be ABI incompatibility.
* ext/mpeg2enc/gstmpeg2enc.cc:
Add Andrew as second maintainer (he's helping me), and also add
an error if no caps was set. This happens if I pull before capsnego
and that's something I should solve sometime else.
* gst/matroska/matroska-demux.c:
(gst_matroska_demux_parse_blockgroup):
Fix time parsing.
* gst/matroska/matroska-mux.c: (gst_matroska_mux_audio_pad_link),
(gst_matroska_mux_track_header):
Add caps to templates.
* gst/mpegaudioparse/gstmpegaudioparse.c: (mp3_sink_factory):
Add mpegversion=1 to prevent confusion with MPEG/AAC.
* gst/mpegstream/gstmpegdemux.c:
Remove layer since it causes warnings about unfixed caps.
* gst/videotestsrc/gstvideotestsrc.c: (gst_videotestsrc_get):
Fix obvious typo (we error out if caps were set, we should of
course error out if *no* caps were set).
* sys/oss/gstosselement.c: (gst_osselement_convert):
Fix format conversion, we confused bits/bytes.
* sys/oss/gstosselement.h:
Improve documentation for 'bps'.
* sys/v4l/TODO:
Remove stuff about plugins that need removing - this was done
ages ago.
* sys/v4l/gstv4lmjpegsrc.c: (gst_v4lmjpegsrc_init),
(gst_v4lmjpegsrc_src_convert), (gst_v4lmjpegsrc_src_query):
* sys/v4l/gstv4lsrc.c: (gst_v4lsrc_init), (gst_v4lsrc_src_convert),
(gst_v4lsrc_src_query):
* sys/v4l2/gstv4l2src.c: (gst_v4l2src_init),
(gst_v4l2src_src_convert), (gst_v4l2src_src_query):
Add get_query_types(), get_formats() and query() functions.
2003-12-21 18:44:34 +00:00
|
|
|
if (audiocontext->bitdepth) {
|
2004-03-15 19:32:27 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_AUDIOBITDEPTH,
|
|
|
|
audiocontext->bitdepth);
|
configure.ac: Improve mpeg2enc detection. This is for distributions that do ship mjpegtools, but without mpeg2enc. Al...
Original commit message from CVS:
2003-12-21 Ronald Bultje <rbultje@ronald.bitfreak.net>
* configure.ac:
Improve mpeg2enc detection. This is for distributions that do
ship mjpegtools, but without mpeg2enc. Also does object check
for might there ever be ABI incompatibility.
* ext/mpeg2enc/gstmpeg2enc.cc:
Add Andrew as second maintainer (he's helping me), and also add
an error if no caps was set. This happens if I pull before capsnego
and that's something I should solve sometime else.
* gst/matroska/matroska-demux.c:
(gst_matroska_demux_parse_blockgroup):
Fix time parsing.
* gst/matroska/matroska-mux.c: (gst_matroska_mux_audio_pad_link),
(gst_matroska_mux_track_header):
Add caps to templates.
* gst/mpegaudioparse/gstmpegaudioparse.c: (mp3_sink_factory):
Add mpegversion=1 to prevent confusion with MPEG/AAC.
* gst/mpegstream/gstmpegdemux.c:
Remove layer since it causes warnings about unfixed caps.
* gst/videotestsrc/gstvideotestsrc.c: (gst_videotestsrc_get):
Fix obvious typo (we error out if caps were set, we should of
course error out if *no* caps were set).
* sys/oss/gstosselement.c: (gst_osselement_convert):
Fix format conversion, we confused bits/bytes.
* sys/oss/gstosselement.h:
Improve documentation for 'bps'.
* sys/v4l/TODO:
Remove stuff about plugins that need removing - this was done
ages ago.
* sys/v4l/gstv4lmjpegsrc.c: (gst_v4lmjpegsrc_init),
(gst_v4lmjpegsrc_src_convert), (gst_v4lmjpegsrc_src_query):
* sys/v4l/gstv4lsrc.c: (gst_v4lsrc_init), (gst_v4lsrc_src_convert),
(gst_v4lsrc_src_query):
* sys/v4l2/gstv4l2src.c: (gst_v4l2src_init),
(gst_v4l2src_src_convert), (gst_v4l2src_src_query):
Add get_query_types(), get_formats() and query() functions.
2003-12-21 18:44:34 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
default:
|
|
|
|
/* doesn't need type-specific data */
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_ebml_write_ascii (ebml, GST_MATROSKA_ID_CODECID, context->codec_id);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (context->codec_priv)
|
|
|
|
gst_ebml_write_binary (ebml, GST_MATROSKA_ID_CODECPRIVATE,
|
2004-03-15 19:32:27 +00:00
|
|
|
context->codec_priv, context->codec_priv_size);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* FIXME: until we have a nice way of getting the codecname
|
|
|
|
* out of the caps, I'm not going to enable this. Too much
|
|
|
|
* (useless, double, boring) work... */
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: Use value from tags if any */
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/*gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_CODECNAME,
|
2004-03-14 22:34:33 +00:00
|
|
|
context->codec_name); */
|
|
|
|
gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_TRACKNAME, context->name);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_start:
|
|
|
|
* @mux: #GstMatroskaMux
|
|
|
|
*
|
|
|
|
* Start a new matroska file (write headers etc...)
|
|
|
|
*/
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_start (GstMatroskaMux * mux)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstEbmlWrite *ebml = mux->ebml_write;
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
guint32 seekhead_id[] = { GST_MATROSKA_ID_SEGMENTINFO,
|
2004-03-14 22:34:33 +00:00
|
|
|
GST_MATROSKA_ID_TRACKS,
|
|
|
|
GST_MATROSKA_ID_CUES,
|
|
|
|
GST_MATROSKA_ID_TAGS,
|
|
|
|
0
|
|
|
|
};
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
guint64 master, child;
|
2005-10-14 12:43:30 +00:00
|
|
|
GSList *collected;
|
|
|
|
int i;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
guint tracknum = 1;
|
2008-03-03 22:01:56 +00:00
|
|
|
GstClockTime duration = 0;
|
2008-08-02 17:47:32 +00:00
|
|
|
guint32 segment_uid[4];
|
2004-09-01 12:10:21 +00:00
|
|
|
GTimeVal time = { 0, 0 };
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* we start with a EBML header */
|
2005-10-28 15:32:48 +00:00
|
|
|
gst_ebml_write_header (ebml, "matroska", mux->matroska_version);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* start a segment */
|
2004-03-14 22:34:33 +00:00
|
|
|
mux->segment_pos =
|
|
|
|
gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_SEGMENT);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
mux->segment_master = ebml->pos;
|
|
|
|
|
|
|
|
/* the rest of the header is cached */
|
|
|
|
gst_ebml_write_set_cache (ebml, 0x1000);
|
|
|
|
|
|
|
|
/* seekhead (table of contents) - we set the positions later */
|
|
|
|
mux->seekhead_pos = ebml->pos;
|
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_SEEKHEAD);
|
|
|
|
for (i = 0; seekhead_id[i] != 0; i++) {
|
|
|
|
child = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_SEEKENTRY);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_SEEKID, seekhead_id[i]);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_SEEKPOSITION, -1);
|
|
|
|
gst_ebml_write_master_finish (ebml, child);
|
|
|
|
}
|
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
|
|
|
|
/* segment info */
|
|
|
|
mux->info_pos = ebml->pos;
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_SEGMENTINFO);
|
2004-09-01 12:10:21 +00:00
|
|
|
for (i = 0; i < 4; i++) {
|
2008-08-02 17:47:32 +00:00
|
|
|
segment_uid[i] = g_random_int ();
|
2004-09-01 12:10:21 +00:00
|
|
|
}
|
|
|
|
gst_ebml_write_binary (ebml, GST_MATROSKA_ID_SEGMENTUID,
|
|
|
|
(guint8 *) segment_uid, 16);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_TIMECODESCALE, mux->time_scale);
|
|
|
|
mux->duration_pos = ebml->pos;
|
2004-08-26 11:10:16 +00:00
|
|
|
/* get duration */
|
2005-10-14 12:43:30 +00:00
|
|
|
for (collected = mux->collect->data; collected;
|
|
|
|
collected = g_slist_next (collected)) {
|
|
|
|
GstMatroskaPad *collect_pad;
|
2005-10-20 19:14:27 +00:00
|
|
|
GstFormat format = GST_FORMAT_TIME;
|
2008-01-08 19:57:23 +00:00
|
|
|
GstPad *thepad;
|
2005-10-20 19:14:27 +00:00
|
|
|
gint64 trackduration;
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
collect_pad = (GstMatroskaPad *) collected->data;
|
|
|
|
thepad = collect_pad->collect.pad;
|
|
|
|
|
|
|
|
/* Query the total length of the track. */
|
2008-01-08 19:57:23 +00:00
|
|
|
GST_DEBUG_OBJECT (thepad, "querying peer duration");
|
|
|
|
if (gst_pad_query_peer_duration (thepad, &format, &trackduration)) {
|
2006-10-05 16:37:33 +00:00
|
|
|
GST_DEBUG_OBJECT (thepad, "duration: %" GST_TIME_FORMAT,
|
2005-11-01 15:11:16 +00:00
|
|
|
GST_TIME_ARGS (trackduration));
|
2008-03-03 22:01:56 +00:00
|
|
|
if (trackduration != GST_CLOCK_TIME_NONE && trackduration > duration) {
|
|
|
|
duration = (GstClockTime) trackduration;
|
2004-08-26 11:10:16 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
gst_ebml_write_float (ebml, GST_MATROSKA_ID_DURATION,
|
2008-03-03 22:01:56 +00:00
|
|
|
gst_guint64_to_gdouble (duration) /
|
|
|
|
gst_guint64_to_gdouble (mux->time_scale));
|
|
|
|
|
2004-08-26 11:10:16 +00:00
|
|
|
gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_MUXINGAPP,
|
2006-04-01 09:56:45 +00:00
|
|
|
"GStreamer plugin version " PACKAGE_VERSION);
|
2005-10-14 12:43:30 +00:00
|
|
|
if (mux->writing_app && mux->writing_app[0]) {
|
|
|
|
gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_WRITINGAPP, mux->writing_app);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
2004-09-01 12:10:21 +00:00
|
|
|
g_get_current_time (&time);
|
|
|
|
gst_ebml_write_date (ebml, GST_MATROSKA_ID_DATEUTC, time.tv_sec);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
|
|
|
|
/* tracks */
|
|
|
|
mux->tracks_pos = ebml->pos;
|
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TRACKS);
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
for (collected = mux->collect->data; collected;
|
|
|
|
collected = g_slist_next (collected)) {
|
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
GstPad *thepad;
|
|
|
|
|
|
|
|
collect_pad = (GstMatroskaPad *) collected->data;
|
|
|
|
thepad = collect_pad->collect.pad;
|
|
|
|
|
2005-11-20 17:04:55 +00:00
|
|
|
if (gst_pad_is_linked (thepad) && gst_pad_is_active (thepad) &&
|
|
|
|
collect_pad->track->codec_id != 0) {
|
2005-10-14 12:43:30 +00:00
|
|
|
collect_pad->track->num = tracknum++;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
child = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TRACKENTRY);
|
2005-10-14 12:43:30 +00:00
|
|
|
gst_matroska_mux_track_header (mux, collect_pad->track);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, child);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
|
|
|
|
/* lastly, flush the cache */
|
|
|
|
gst_ebml_write_flush_cache (ebml);
|
|
|
|
}
|
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
static void
|
|
|
|
gst_matroska_mux_write_simple_tag (const GstTagList * list, const gchar * tag,
|
|
|
|
gpointer data)
|
|
|
|
{
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: more sensible tag mappings */
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
struct
|
|
|
|
{
|
|
|
|
gchar *matroska_tagname;
|
|
|
|
gchar *gstreamer_tagname;
|
|
|
|
}
|
|
|
|
tag_conv[] = {
|
|
|
|
{
|
|
|
|
GST_MATROSKA_TAG_ID_TITLE, GST_TAG_TITLE}, {
|
|
|
|
GST_MATROSKA_TAG_ID_AUTHOR, GST_TAG_ARTIST}, {
|
|
|
|
GST_MATROSKA_TAG_ID_ALBUM, GST_TAG_ALBUM}, {
|
|
|
|
GST_MATROSKA_TAG_ID_COMMENTS, GST_TAG_COMMENT}, {
|
|
|
|
GST_MATROSKA_TAG_ID_BITSPS, GST_TAG_BITRATE}, {
|
2008-06-11 11:18:23 +00:00
|
|
|
GST_MATROSKA_TAG_ID_BPS, GST_TAG_BITRATE}, {
|
|
|
|
GST_MATROSKA_TAG_ID_ENCODER, GST_TAG_ENCODER}, {
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
GST_MATROSKA_TAG_ID_DATE, GST_TAG_DATE}, {
|
|
|
|
GST_MATROSKA_TAG_ID_ISRC, GST_TAG_ISRC}, {
|
2008-06-11 11:18:23 +00:00
|
|
|
GST_MATROSKA_TAG_ID_COPYRIGHT, GST_TAG_COPYRIGHT}, {
|
|
|
|
GST_MATROSKA_TAG_ID_BPM, GST_TAG_BEATS_PER_MINUTE}, {
|
|
|
|
GST_MATROSKA_TAG_ID_TERMS_OF_USE, GST_TAG_LICENSE}, {
|
|
|
|
GST_MATROSKA_TAG_ID_COMPOSER, GST_TAG_COMPOSER}, {
|
|
|
|
GST_MATROSKA_TAG_ID_LEAD_PERFORMER, GST_TAG_PERFORMER}, {
|
|
|
|
GST_MATROSKA_TAG_ID_GENRE, GST_TAG_GENRE}
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
};
|
|
|
|
GstEbmlWrite *ebml = (GstEbmlWrite *) data;
|
|
|
|
guint i;
|
|
|
|
guint64 simpletag_master;
|
|
|
|
|
|
|
|
for (i = 0; i < G_N_ELEMENTS (tag_conv); i++) {
|
|
|
|
const gchar *tagname_gst = tag_conv[i].gstreamer_tagname;
|
|
|
|
const gchar *tagname_mkv = tag_conv[i].matroska_tagname;
|
|
|
|
|
|
|
|
if (strcmp (tagname_gst, tag) == 0) {
|
|
|
|
GValue src = { 0, };
|
2008-06-13 19:14:41 +00:00
|
|
|
gchar *dest;
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
|
|
|
|
if (!gst_tag_list_copy_value (&src, list, tag))
|
|
|
|
break;
|
2008-06-13 19:14:41 +00:00
|
|
|
if ((dest = gst_value_serialize (&src))) {
|
2008-06-11 11:18:23 +00:00
|
|
|
|
|
|
|
simpletag_master = gst_ebml_write_master_start (ebml,
|
|
|
|
GST_MATROSKA_ID_SIMPLETAG);
|
|
|
|
gst_ebml_write_ascii (ebml, GST_MATROSKA_ID_TAGNAME, tagname_mkv);
|
2008-06-13 19:14:41 +00:00
|
|
|
gst_ebml_write_utf8 (ebml, GST_MATROSKA_ID_TAGSTRING, dest);
|
2008-06-11 11:18:23 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, simpletag_master);
|
2009-03-31 14:16:04 +00:00
|
|
|
g_free (dest);
|
2008-06-11 11:18:23 +00:00
|
|
|
} else {
|
|
|
|
GST_WARNING ("Can't transform tag '%s' to string", tagname_mkv);
|
|
|
|
}
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
g_value_unset (&src);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_finish:
|
|
|
|
* @mux: #GstMatroskaMux
|
|
|
|
*
|
|
|
|
* Finish a new matroska file (write index etc...)
|
|
|
|
*/
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_finish (GstMatroskaMux * mux)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstEbmlWrite *ebml = mux->ebml_write;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
guint64 pos;
|
2004-08-26 11:10:16 +00:00
|
|
|
guint64 duration = 0;
|
2005-10-14 12:43:30 +00:00
|
|
|
GSList *collected;
|
2009-01-06 17:48:10 +00:00
|
|
|
const GstTagList *tags;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
/* finish last cluster */
|
|
|
|
if (mux->cluster) {
|
|
|
|
gst_ebml_write_master_finish (ebml, mux->cluster);
|
|
|
|
}
|
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* cues */
|
|
|
|
if (mux->index != NULL) {
|
|
|
|
guint n;
|
|
|
|
guint64 master, pointentry_master, trackpos_master;
|
|
|
|
|
|
|
|
mux->cues_pos = ebml->pos;
|
|
|
|
gst_ebml_write_set_cache (ebml, 12 + 41 * mux->num_indexes);
|
|
|
|
master = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_CUES);
|
|
|
|
|
|
|
|
for (n = 0; n < mux->num_indexes; n++) {
|
|
|
|
GstMatroskaIndex *idx = &mux->index[n];
|
|
|
|
|
|
|
|
pointentry_master = gst_ebml_write_master_start (ebml,
|
2004-03-15 19:32:27 +00:00
|
|
|
GST_MATROSKA_ID_POINTENTRY);
|
2004-09-01 13:41:37 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_CUETIME,
|
2004-03-15 19:32:27 +00:00
|
|
|
idx->time / mux->time_scale);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
trackpos_master = gst_ebml_write_master_start (ebml,
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
GST_MATROSKA_ID_CUETRACKPOSITIONS);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_CUETRACK, idx->track);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_CUECLUSTERPOSITION,
|
2004-03-15 19:32:27 +00:00
|
|
|
idx->pos - mux->segment_master);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, trackpos_master);
|
|
|
|
gst_ebml_write_master_finish (ebml, pointentry_master);
|
|
|
|
}
|
|
|
|
|
|
|
|
gst_ebml_write_master_finish (ebml, master);
|
|
|
|
gst_ebml_write_flush_cache (ebml);
|
|
|
|
}
|
|
|
|
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/* tags */
|
2009-01-06 17:48:10 +00:00
|
|
|
tags = gst_tag_setter_get_tag_list (GST_TAG_SETTER (mux));
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
|
2009-01-06 17:48:10 +00:00
|
|
|
if (tags != NULL && !gst_tag_list_is_empty (tags)) {
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
guint64 master_tags, master_tag;
|
|
|
|
|
2008-08-25 14:15:43 +00:00
|
|
|
GST_DEBUG ("Writing tags");
|
|
|
|
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: maybe limit via the TARGETS id by looking at the source pad */
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
mux->tags_pos = ebml->pos;
|
|
|
|
master_tags = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TAGS);
|
|
|
|
master_tag = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_TAG);
|
|
|
|
gst_tag_list_foreach (tags, gst_matroska_mux_write_simple_tag, ebml);
|
|
|
|
gst_ebml_write_master_finish (ebml, master_tag);
|
|
|
|
gst_ebml_write_master_finish (ebml, master_tags);
|
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* update seekhead. We know that:
|
|
|
|
* - a seekhead contains 4 entries.
|
|
|
|
* - order of entries is as above.
|
|
|
|
* - a seekhead has a 4-byte header + 8-byte length
|
|
|
|
* - each entry is 2-byte master, 2-byte ID pointer,
|
|
|
|
* 2-byte length pointer, all 8/1-byte length, 4-
|
|
|
|
* byte ID and 8-byte length pointer, where the
|
|
|
|
* length pointer starts at 20.
|
|
|
|
* - all entries are local to the segment (so pos - segment_master).
|
|
|
|
* - so each entry is at 12 + 20 + num * 28. */
|
|
|
|
gst_ebml_replace_uint (ebml, mux->seekhead_pos + 32,
|
2004-03-14 22:34:33 +00:00
|
|
|
mux->info_pos - mux->segment_master);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_replace_uint (ebml, mux->seekhead_pos + 60,
|
2004-03-14 22:34:33 +00:00
|
|
|
mux->tracks_pos - mux->segment_master);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
if (mux->index != NULL) {
|
|
|
|
gst_ebml_replace_uint (ebml, mux->seekhead_pos + 88,
|
2004-03-15 19:32:27 +00:00
|
|
|
mux->cues_pos - mux->segment_master);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
} else {
|
|
|
|
/* void'ify */
|
|
|
|
guint64 my_pos = ebml->pos;
|
2004-03-14 22:34:33 +00:00
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_ebml_write_seek (ebml, mux->seekhead_pos + 68);
|
|
|
|
gst_ebml_write_buffer_header (ebml, GST_EBML_ID_VOID, 26);
|
|
|
|
gst_ebml_write_seek (ebml, my_pos);
|
|
|
|
}
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
if (tags != NULL) {
|
2008-06-15 15:29:29 +00:00
|
|
|
gst_ebml_replace_uint (ebml, mux->seekhead_pos + 116,
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
mux->tags_pos - mux->segment_master);
|
|
|
|
} else {
|
|
|
|
/* void'ify */
|
|
|
|
guint64 my_pos = ebml->pos;
|
|
|
|
|
2008-06-15 15:29:29 +00:00
|
|
|
gst_ebml_write_seek (ebml, mux->seekhead_pos + 96);
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
gst_ebml_write_buffer_header (ebml, GST_EBML_ID_VOID, 26);
|
|
|
|
gst_ebml_write_seek (ebml, my_pos);
|
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* update duration */
|
2004-08-26 11:10:16 +00:00
|
|
|
/* first get the overall duration */
|
gst/matroska/matroska-mux.*: Add basic tag writing support; implement releasing pads (#374658).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/matroska/matroska-mux.c: (gst_matroska_mux_add_interfaces),
(gst_matroska_mux_class_init), (gst_matroska_pad_free),
(gst_matroska_mux_reset), (gst_matroska_mux_handle_sink_event),
(gst_matroska_mux_request_new_pad), (gst_matroska_mux_release_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish):
* gst/matroska/matroska-mux.h:
Add basic tag writing support; implement releasing pads (#374658).
2006-11-15 12:08:20 +00:00
|
|
|
/* a released track may have left a duration in here */
|
|
|
|
duration = mux->duration;
|
2005-10-14 12:43:30 +00:00
|
|
|
for (collected = mux->collect->data; collected;
|
|
|
|
collected = g_slist_next (collected)) {
|
|
|
|
GstMatroskaPad *collect_pad;
|
2008-01-08 19:57:23 +00:00
|
|
|
GstClockTime min_duration; /* observed minimum duration */
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
collect_pad = (GstMatroskaPad *) collected->data;
|
|
|
|
|
2008-03-03 22:01:56 +00:00
|
|
|
GST_DEBUG_OBJECT (mux, "Pad %" GST_PTR_FORMAT " start ts %" GST_TIME_FORMAT
|
|
|
|
" end ts %" GST_TIME_FORMAT, collect_pad,
|
|
|
|
GST_TIME_ARGS (collect_pad->start_ts),
|
|
|
|
GST_TIME_ARGS (collect_pad->end_ts));
|
|
|
|
|
|
|
|
if (GST_CLOCK_TIME_IS_VALID (collect_pad->start_ts) &&
|
|
|
|
GST_CLOCK_TIME_IS_VALID (collect_pad->end_ts)) {
|
|
|
|
min_duration =
|
|
|
|
GST_CLOCK_DIFF (collect_pad->start_ts, collect_pad->end_ts);
|
|
|
|
if (collect_pad->duration < min_duration)
|
|
|
|
collect_pad->duration = min_duration;
|
|
|
|
GST_DEBUG_OBJECT (collect_pad, "final track duration: %" GST_TIME_FORMAT,
|
|
|
|
GST_TIME_ARGS (collect_pad->duration));
|
|
|
|
}
|
|
|
|
|
|
|
|
if (GST_CLOCK_TIME_IS_VALID (collect_pad->duration) &&
|
|
|
|
duration < collect_pad->duration)
|
2005-10-14 12:43:30 +00:00
|
|
|
duration = collect_pad->duration;
|
2004-08-26 11:10:16 +00:00
|
|
|
}
|
|
|
|
if (duration != 0) {
|
2008-01-08 19:57:23 +00:00
|
|
|
GST_DEBUG_OBJECT (mux, "final total duration: %" GST_TIME_FORMAT,
|
|
|
|
GST_TIME_ARGS (duration));
|
2005-10-14 12:43:30 +00:00
|
|
|
pos = mux->ebml_write->pos;
|
2004-08-26 11:10:16 +00:00
|
|
|
gst_ebml_write_seek (ebml, mux->duration_pos);
|
|
|
|
gst_ebml_write_float (ebml, GST_MATROSKA_ID_DURATION,
|
2008-03-03 22:01:56 +00:00
|
|
|
gst_guint64_to_gdouble (duration) /
|
|
|
|
gst_guint64_to_gdouble (mux->time_scale));
|
2004-08-26 11:10:16 +00:00
|
|
|
gst_ebml_write_seek (ebml, pos);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
} else {
|
|
|
|
/* void'ify */
|
|
|
|
guint64 my_pos = ebml->pos;
|
|
|
|
|
|
|
|
gst_ebml_write_seek (ebml, mux->duration_pos);
|
|
|
|
gst_ebml_write_buffer_header (ebml, GST_EBML_ID_VOID, 8);
|
|
|
|
gst_ebml_write_seek (ebml, my_pos);
|
2004-08-26 11:10:16 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* finish segment - this also writes element length */
|
|
|
|
gst_ebml_write_master_finish (ebml, mux->segment_pos);
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_best_pad:
|
|
|
|
* @mux: #GstMatroskaMux
|
2006-01-23 10:44:03 +00:00
|
|
|
* @popped: True if at least one buffer was popped from #GstCollectPads
|
2005-10-14 12:43:30 +00:00
|
|
|
*
|
|
|
|
* Find a pad with the oldest data
|
|
|
|
* (data from this pad should be written first).
|
|
|
|
*
|
|
|
|
* Returns: Selected pad.
|
|
|
|
*/
|
|
|
|
static GstMatroskaPad *
|
2006-01-23 10:44:03 +00:00
|
|
|
gst_matroska_mux_best_pad (GstMatroskaMux * mux, gboolean * popped)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GSList *collected;
|
|
|
|
GstMatroskaPad *best = NULL;
|
|
|
|
|
2006-01-23 10:44:03 +00:00
|
|
|
*popped = FALSE;
|
2005-10-14 12:43:30 +00:00
|
|
|
for (collected = mux->collect->data; collected;
|
|
|
|
collected = g_slist_next (collected)) {
|
|
|
|
GstMatroskaPad *collect_pad;
|
|
|
|
|
|
|
|
collect_pad = (GstMatroskaPad *) collected->data;
|
|
|
|
/* fetch a new buffer if needed */
|
|
|
|
if (collect_pad->buffer == NULL) {
|
Fix for collect pads API change. Also fix textoverlay state change function.
Original commit message from CVS:
* ext/cairo/gsttextoverlay.c: (gst_text_overlay_finalize),
(gst_text_overlay_init), (gst_text_overlay_text_pad_linked),
(gst_text_overlay_text_pad_unlinked), (gst_text_overlay_pop_video),
(gst_text_overlay_pop_text), (gst_text_overlay_collected),
(gst_text_overlay_change_state):
* gst/matroska/matroska-mux.c: (gst_matroska_mux_init),
(gst_matroska_mux_reset), (gst_matroska_mux_request_new_pad),
(gst_matroska_mux_best_pad), (gst_matroska_mux_change_state):
* gst/smpte/gstsmpte.c: (gst_smpte_init), (gst_smpte_collected):
* gst/videomixer/videomixer.c: (gst_videomixer_init),
(gst_videomixer_request_new_pad), (gst_videomixer_fill_queues),
(gst_videomixer_change_state):
Fix for collect pads API change. Also fix textoverlay state
change function.
2005-11-21 12:13:48 +00:00
|
|
|
collect_pad->buffer = gst_collect_pads_pop (mux->collect,
|
2005-10-14 12:43:30 +00:00
|
|
|
(GstCollectData *) collect_pad);
|
2006-01-23 10:44:03 +00:00
|
|
|
|
|
|
|
if (collect_pad->buffer != NULL)
|
|
|
|
*popped = TRUE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/* if we have a buffer check if it is better then the current best one */
|
|
|
|
if (collect_pad->buffer != NULL) {
|
2006-05-03 14:51:50 +00:00
|
|
|
if (best == NULL || !GST_BUFFER_TIMESTAMP_IS_VALID (collect_pad->buffer)
|
|
|
|
|| (GST_BUFFER_TIMESTAMP_IS_VALID (best->buffer)
|
|
|
|
&& GST_BUFFER_TIMESTAMP (collect_pad->buffer) <
|
|
|
|
GST_BUFFER_TIMESTAMP (best->buffer))) {
|
2005-10-14 12:43:30 +00:00
|
|
|
best = collect_pad;
|
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return best;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-28 15:32:48 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_buffer_header:
|
|
|
|
* @track: Track context.
|
|
|
|
* @relative_timestamp: relative timestamp of the buffer
|
|
|
|
* @flags: Buffer flags.
|
|
|
|
*
|
|
|
|
* Create a buffer containing buffer header.
|
|
|
|
*
|
|
|
|
* Returns: New buffer.
|
|
|
|
*/
|
|
|
|
GstBuffer *
|
|
|
|
gst_matroska_mux_create_buffer_header (GstMatroskaTrackContext * track,
|
2006-04-25 11:09:24 +00:00
|
|
|
gint16 relative_timestamp, int flags)
|
2005-10-28 15:32:48 +00:00
|
|
|
{
|
|
|
|
GstBuffer *hdr;
|
|
|
|
|
|
|
|
hdr = gst_buffer_new_and_alloc (4);
|
|
|
|
/* track num - FIXME: what if num >= 0x80 (unlikely)? */
|
|
|
|
GST_BUFFER_DATA (hdr)[0] = track->num | 0x80;
|
|
|
|
/* time relative to clustertime */
|
|
|
|
GST_WRITE_UINT16_BE (GST_BUFFER_DATA (hdr) + 1, relative_timestamp);
|
|
|
|
|
|
|
|
/* flags */
|
|
|
|
GST_BUFFER_DATA (hdr)[3] = flags;
|
|
|
|
|
|
|
|
return hdr;
|
|
|
|
}
|
|
|
|
|
2009-09-22 10:48:50 +00:00
|
|
|
#define DIRAC_PARSE_CODE_SEQUENCE 0x00
|
|
|
|
|
2008-11-05 14:42:35 +00:00
|
|
|
static GstBuffer *
|
|
|
|
gst_matroska_mux_handle_dirac_packet (GstMatroskaMux * mux,
|
|
|
|
GstMatroskaPad * collect_pad, GstBuffer * buf)
|
|
|
|
{
|
|
|
|
GstMatroskaTrackVideoContext *ctx =
|
|
|
|
(GstMatroskaTrackVideoContext *) collect_pad->track;
|
|
|
|
const guint8 *data = GST_BUFFER_DATA (buf);
|
2008-11-11 12:18:23 +00:00
|
|
|
guint size = GST_BUFFER_SIZE (buf);
|
2008-11-05 14:42:35 +00:00
|
|
|
guint8 parse_code;
|
2008-11-11 12:18:23 +00:00
|
|
|
guint32 next_parse_offset;
|
2008-11-05 14:42:35 +00:00
|
|
|
GstBuffer *ret = NULL;
|
2008-11-11 12:18:23 +00:00
|
|
|
gboolean is_picture = FALSE;
|
2008-11-05 14:42:35 +00:00
|
|
|
|
|
|
|
if (GST_BUFFER_SIZE (buf) < 13) {
|
|
|
|
gst_buffer_unref (buf);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-11-11 12:18:23 +00:00
|
|
|
/* Check if this buffer contains a picture packet */
|
|
|
|
while (size >= 13) {
|
2009-09-22 10:48:50 +00:00
|
|
|
if (GST_READ_UINT32_BE (data) != 0x42424344 /* 'BBCD' */ ) {
|
2008-11-11 12:18:23 +00:00
|
|
|
gst_buffer_unref (buf);
|
|
|
|
return ret;
|
|
|
|
}
|
2008-11-05 14:42:35 +00:00
|
|
|
|
2008-11-11 12:18:23 +00:00
|
|
|
parse_code = GST_READ_UINT8 (data + 4);
|
2009-09-22 10:48:50 +00:00
|
|
|
if (parse_code == DIRAC_PARSE_CODE_SEQUENCE) {
|
2008-11-05 14:42:35 +00:00
|
|
|
if (ctx->dirac_unit) {
|
|
|
|
gst_buffer_unref (ctx->dirac_unit);
|
|
|
|
ctx->dirac_unit = NULL;
|
|
|
|
}
|
2008-11-11 12:18:23 +00:00
|
|
|
} else if (parse_code & 0x08) {
|
|
|
|
is_picture = TRUE;
|
2008-11-05 14:42:35 +00:00
|
|
|
break;
|
2008-11-11 12:18:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
next_parse_offset = GST_READ_UINT32_BE (data + 5);
|
|
|
|
|
2009-09-22 10:48:50 +00:00
|
|
|
if (G_UNLIKELY (next_parse_offset == 0))
|
|
|
|
break;
|
|
|
|
|
2008-11-11 12:18:23 +00:00
|
|
|
data += next_parse_offset;
|
|
|
|
size -= next_parse_offset;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ctx->dirac_unit)
|
|
|
|
ctx->dirac_unit = gst_buffer_join (ctx->dirac_unit, gst_buffer_ref (buf));
|
|
|
|
else
|
|
|
|
ctx->dirac_unit = gst_buffer_ref (buf);
|
|
|
|
|
|
|
|
if (is_picture) {
|
|
|
|
ret = gst_buffer_make_metadata_writable (ctx->dirac_unit);
|
|
|
|
ctx->dirac_unit = NULL;
|
|
|
|
gst_buffer_copy_metadata (ret, buf,
|
|
|
|
GST_BUFFER_COPY_FLAGS | GST_BUFFER_COPY_TIMESTAMPS |
|
|
|
|
GST_BUFFER_COPY_CAPS);
|
|
|
|
gst_buffer_unref (buf);
|
|
|
|
} else {
|
|
|
|
gst_buffer_unref (buf);
|
|
|
|
ret = NULL;
|
2008-11-05 14:42:35 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
/**
|
|
|
|
* gst_matroska_mux_write_data:
|
|
|
|
* @mux: #GstMatroskaMux
|
2006-01-23 10:44:03 +00:00
|
|
|
* @collect_pad: #GstMatroskaPad with the data
|
2005-10-14 12:43:30 +00:00
|
|
|
*
|
|
|
|
* Write collected data (called from gst_matroska_mux_collected).
|
|
|
|
*
|
|
|
|
* Returns: Result of the gst_pad_push issued to write the data.
|
|
|
|
*/
|
|
|
|
static GstFlowReturn
|
2006-01-23 10:44:03 +00:00
|
|
|
gst_matroska_mux_write_data (GstMatroskaMux * mux, GstMatroskaPad * collect_pad)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstEbmlWrite *ebml = mux->ebml_write;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstBuffer *buf, *hdr;
|
2009-04-18 16:11:00 +00:00
|
|
|
guint64 blockgroup;
|
2005-10-28 15:32:48 +00:00
|
|
|
gboolean write_duration;
|
2006-04-25 11:09:24 +00:00
|
|
|
gint16 relative_timestamp;
|
|
|
|
gint64 relative_timestamp64;
|
2005-11-10 12:22:30 +00:00
|
|
|
guint64 block_duration;
|
2006-05-03 18:41:47 +00:00
|
|
|
gboolean is_video_keyframe = FALSE;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
|
|
|
/* write data */
|
2006-01-23 10:44:03 +00:00
|
|
|
buf = collect_pad->buffer;
|
|
|
|
collect_pad->buffer = NULL;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
/* vorbis/theora headers are retrieved from caps and put in CodecPrivate */
|
|
|
|
if (collect_pad->track->xiph_headers_to_skip > 0) {
|
|
|
|
GST_LOG_OBJECT (collect_pad->collect.pad, "dropping streamheader buffer");
|
2006-05-03 14:51:50 +00:00
|
|
|
gst_buffer_unref (buf);
|
gst/matroska/: Add support for muxing/demuxing theora video (#342448; too bad none of the usual linux players can act...
Original commit message from CVS:
* gst/matroska/matroska-demux.c: (gst_matroska_demux_add_stream),
(gst_matroska_demux_push_xiph_codec_priv_data),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_video_caps), (gst_matroska_demux_audio_caps):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_video_pad_setcaps),
(xiph3_streamheader_to_codecdata),
(vorbis_streamheader_to_codecdata),
(theora_streamheader_to_codecdata),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_write_data):
Add support for muxing/demuxing theora video (#342448; too bad
none of the usual linux players can actually play this). Playback
in GStreamer will require additional changes to theoradec in -base.
Refactor streamheaders <=> CodecPrivateData code a bit; some small
cleanups.
2006-05-23 13:44:11 +00:00
|
|
|
--collect_pad->track->xiph_headers_to_skip;
|
2006-05-03 14:51:50 +00:00
|
|
|
return GST_FLOW_OK;
|
|
|
|
}
|
|
|
|
|
2008-11-05 14:42:35 +00:00
|
|
|
/* for dirac we have to queue up everything up to a picture unit */
|
2008-11-20 14:30:40 +00:00
|
|
|
if (collect_pad->track->codec_id != NULL &&
|
|
|
|
strcmp (collect_pad->track->codec_id,
|
2008-11-05 14:42:35 +00:00
|
|
|
GST_MATROSKA_CODEC_ID_VIDEO_DIRAC) == 0) {
|
|
|
|
buf = gst_matroska_mux_handle_dirac_packet (mux, collect_pad, buf);
|
|
|
|
if (!buf)
|
|
|
|
return GST_FLOW_OK;
|
|
|
|
}
|
|
|
|
|
2006-05-03 14:51:50 +00:00
|
|
|
/* hm, invalid timestamp (due to --to be fixed--- element upstream);
|
|
|
|
* this would wreak havoc with time stored in matroska file */
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
/* TODO: maybe calculate a timestamp by using the previous timestamp
|
|
|
|
* and default duration */
|
2006-05-03 14:51:50 +00:00
|
|
|
if (!GST_BUFFER_TIMESTAMP_IS_VALID (buf)) {
|
|
|
|
GST_WARNING_OBJECT (collect_pad->collect.pad,
|
|
|
|
"Invalid buffer timestamp; dropping buffer");
|
|
|
|
gst_buffer_unref (buf);
|
|
|
|
return GST_FLOW_OK;
|
|
|
|
}
|
|
|
|
|
2005-11-01 12:35:39 +00:00
|
|
|
/* set the timestamp for outgoing buffers */
|
|
|
|
ebml->timestamp = GST_BUFFER_TIMESTAMP (buf);
|
|
|
|
|
2006-05-03 18:41:47 +00:00
|
|
|
if (collect_pad->track->type == GST_MATROSKA_TRACK_TYPE_VIDEO &&
|
|
|
|
!GST_BUFFER_FLAG_IS_SET (buf, GST_BUFFER_FLAG_DELTA_UNIT)) {
|
|
|
|
GST_LOG_OBJECT (mux, "have video keyframe, ts=%" GST_TIME_FORMAT,
|
|
|
|
GST_TIME_ARGS (GST_BUFFER_TIMESTAMP (buf)));
|
|
|
|
is_video_keyframe = TRUE;
|
|
|
|
}
|
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
if (mux->cluster) {
|
2006-05-03 14:51:50 +00:00
|
|
|
/* start a new cluster every two seconds or at keyframe */
|
|
|
|
if (mux->cluster_time + GST_SECOND * 2 < GST_BUFFER_TIMESTAMP (buf)
|
2006-05-03 18:41:47 +00:00
|
|
|
|| is_video_keyframe) {
|
2004-09-09 17:54:26 +00:00
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, mux->cluster);
|
|
|
|
mux->cluster_pos = ebml->pos;
|
|
|
|
mux->cluster =
|
|
|
|
gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_CLUSTER);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_CLUSTERTIMECODE,
|
|
|
|
GST_BUFFER_TIMESTAMP (buf) / mux->time_scale);
|
|
|
|
mux->cluster_time = GST_BUFFER_TIMESTAMP (buf);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* first cluster */
|
2004-09-09 17:54:26 +00:00
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
mux->cluster_pos = ebml->pos;
|
|
|
|
mux->cluster = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_CLUSTER);
|
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_CLUSTERTIMECODE,
|
|
|
|
GST_BUFFER_TIMESTAMP (buf) / mux->time_scale);
|
|
|
|
mux->cluster_time = GST_BUFFER_TIMESTAMP (buf);
|
|
|
|
}
|
|
|
|
|
2004-08-26 11:10:16 +00:00
|
|
|
/* update duration of this track */
|
|
|
|
if (GST_BUFFER_DURATION_IS_VALID (buf))
|
2006-01-23 10:44:03 +00:00
|
|
|
collect_pad->duration += GST_BUFFER_DURATION (buf);
|
2004-08-26 11:10:16 +00:00
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* We currently write an index entry for each keyframe in a
|
2004-09-01 13:41:37 +00:00
|
|
|
* video track or one entry for each cluster in an audio track
|
|
|
|
* for audio only files. This can be largely improved, such as doing
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
* one for each keyframe or each second (for all-keyframe
|
2004-09-01 13:41:37 +00:00
|
|
|
* streams), only the *first* video track. But that'll come later... */
|
2008-06-11 11:27:46 +00:00
|
|
|
|
2008-06-13 19:14:41 +00:00
|
|
|
/* TODO: index is useful for every track, should contain the number of
|
|
|
|
* the block in the cluster which contains the timestamp
|
|
|
|
*/
|
2006-05-03 18:41:47 +00:00
|
|
|
if (is_video_keyframe) {
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstMatroskaIndex *idx;
|
|
|
|
|
|
|
|
if (mux->num_indexes % 32 == 0) {
|
|
|
|
mux->index = g_renew (GstMatroskaIndex, mux->index,
|
2004-03-15 19:32:27 +00:00
|
|
|
mux->num_indexes + 32);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
idx = &mux->index[mux->num_indexes++];
|
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
idx->pos = mux->cluster_pos;
|
2004-09-01 13:41:37 +00:00
|
|
|
idx->time = GST_BUFFER_TIMESTAMP (buf);
|
2006-01-23 10:44:03 +00:00
|
|
|
idx->track = collect_pad->track->num;
|
|
|
|
} else if ((collect_pad->track->type == GST_MATROSKA_TRACK_TYPE_AUDIO) &&
|
2004-09-01 13:41:37 +00:00
|
|
|
(mux->num_streams == 1)) {
|
|
|
|
GstMatroskaIndex *idx;
|
|
|
|
|
|
|
|
if (mux->num_indexes % 32 == 0) {
|
|
|
|
mux->index = g_renew (GstMatroskaIndex, mux->index,
|
|
|
|
mux->num_indexes + 32);
|
|
|
|
}
|
|
|
|
idx = &mux->index[mux->num_indexes++];
|
|
|
|
|
2004-09-09 15:40:06 +00:00
|
|
|
idx->pos = mux->cluster_pos;
|
2004-03-14 22:34:33 +00:00
|
|
|
idx->time = GST_BUFFER_TIMESTAMP (buf);
|
2006-01-23 10:44:03 +00:00
|
|
|
idx->track = collect_pad->track->num;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-28 15:32:48 +00:00
|
|
|
/* Check if the duration differs from the default duration. */
|
|
|
|
write_duration = FALSE;
|
2005-11-10 12:22:30 +00:00
|
|
|
block_duration = GST_BUFFER_DURATION (buf);
|
2004-09-01 12:10:21 +00:00
|
|
|
if (GST_BUFFER_DURATION_IS_VALID (buf)) {
|
2006-01-23 10:44:03 +00:00
|
|
|
if (block_duration != collect_pad->track->default_duration) {
|
2005-10-28 15:32:48 +00:00
|
|
|
write_duration = TRUE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* write the block, for matroska v2 use SimpleBlock if possible
|
|
|
|
* one slice (*breath*).
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
* FIXME: Need to do correct lacing! */
|
2006-04-25 11:09:24 +00:00
|
|
|
relative_timestamp64 = GST_BUFFER_TIMESTAMP (buf) - mux->cluster_time;
|
|
|
|
if (relative_timestamp64 >= 0) {
|
|
|
|
/* round the timestamp */
|
|
|
|
relative_timestamp64 += mux->time_scale / 2;
|
|
|
|
} else {
|
|
|
|
/* round the timestamp */
|
|
|
|
relative_timestamp64 -= mux->time_scale / 2;
|
|
|
|
}
|
|
|
|
relative_timestamp = relative_timestamp64 / (gint64) mux->time_scale;
|
2005-10-28 15:32:48 +00:00
|
|
|
if (mux->matroska_version > 1 && !write_duration) {
|
|
|
|
int flags =
|
|
|
|
GST_BUFFER_FLAG_IS_SET (buf, GST_BUFFER_FLAG_DELTA_UNIT) ? 0 : 0x80;
|
|
|
|
|
|
|
|
hdr =
|
2006-01-23 10:44:03 +00:00
|
|
|
gst_matroska_mux_create_buffer_header (collect_pad->track,
|
|
|
|
relative_timestamp, flags);
|
2005-10-28 15:32:48 +00:00
|
|
|
gst_ebml_write_buffer_header (ebml, GST_MATROSKA_ID_SIMPLEBLOCK,
|
|
|
|
GST_BUFFER_SIZE (buf) + GST_BUFFER_SIZE (hdr));
|
|
|
|
gst_ebml_write_buffer (ebml, hdr);
|
|
|
|
gst_ebml_write_buffer (ebml, buf);
|
|
|
|
|
|
|
|
return gst_ebml_last_write_result (ebml);
|
|
|
|
} else {
|
|
|
|
blockgroup = gst_ebml_write_master_start (ebml, GST_MATROSKA_ID_BLOCKGROUP);
|
|
|
|
hdr =
|
2006-01-23 10:44:03 +00:00
|
|
|
gst_matroska_mux_create_buffer_header (collect_pad->track,
|
|
|
|
relative_timestamp, 0);
|
2005-10-28 15:32:48 +00:00
|
|
|
gst_ebml_write_buffer_header (ebml, GST_MATROSKA_ID_BLOCK,
|
|
|
|
GST_BUFFER_SIZE (buf) + GST_BUFFER_SIZE (hdr));
|
|
|
|
gst_ebml_write_buffer (ebml, hdr);
|
|
|
|
gst_ebml_write_buffer (ebml, buf);
|
|
|
|
if (write_duration) {
|
2004-09-01 12:10:21 +00:00
|
|
|
gst_ebml_write_uint (ebml, GST_MATROSKA_ID_BLOCKDURATION,
|
|
|
|
block_duration / mux->time_scale);
|
|
|
|
}
|
2005-10-28 15:32:48 +00:00
|
|
|
gst_ebml_write_master_finish (ebml, blockgroup);
|
|
|
|
return gst_ebml_last_write_result (ebml);
|
2004-09-01 12:10:21 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_collected:
|
|
|
|
* @pads: #GstCollectPads
|
|
|
|
* @uuser_data: #GstMatroskaMux
|
|
|
|
*
|
|
|
|
* Collectpads callback.
|
|
|
|
*
|
|
|
|
* Returns: #GstFlowReturn
|
|
|
|
*/
|
|
|
|
static GstFlowReturn
|
|
|
|
gst_matroska_mux_collected (GstCollectPads * pads, gpointer user_data)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstMatroskaMux *mux = GST_MATROSKA_MUX (user_data);
|
2006-01-23 10:44:03 +00:00
|
|
|
GstMatroskaPad *best;
|
|
|
|
gboolean popped;
|
|
|
|
GstFlowReturn ret;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
|
2005-11-01 15:11:16 +00:00
|
|
|
GST_DEBUG_OBJECT (mux, "Collected pads");
|
2005-01-09 10:30:59 +00:00
|
|
|
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
/* start with a header */
|
|
|
|
if (mux->state == GST_MATROSKA_MUX_STATE_START) {
|
2005-10-14 12:43:30 +00:00
|
|
|
if (mux->collect->data == NULL) {
|
|
|
|
GST_ELEMENT_ERROR (mux, STREAM, MUX, (NULL),
|
|
|
|
("No input streams configured"));
|
|
|
|
return GST_FLOW_ERROR;
|
2004-08-26 11:10:16 +00:00
|
|
|
}
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
mux->state = GST_MATROSKA_MUX_STATE_HEADER;
|
|
|
|
gst_matroska_mux_start (mux);
|
|
|
|
mux->state = GST_MATROSKA_MUX_STATE_DATA;
|
|
|
|
}
|
|
|
|
|
2006-01-23 10:44:03 +00:00
|
|
|
do {
|
|
|
|
/* which stream to write from? */
|
|
|
|
best = gst_matroska_mux_best_pad (mux, &popped);
|
|
|
|
|
|
|
|
/* if there is no best pad, we have reached EOS */
|
|
|
|
if (best == NULL) {
|
|
|
|
GST_DEBUG_OBJECT (mux, "No best pad finishing...");
|
|
|
|
gst_matroska_mux_finish (mux);
|
|
|
|
gst_pad_push_event (mux->srcpad, gst_event_new_eos ());
|
|
|
|
ret = GST_FLOW_UNEXPECTED;
|
|
|
|
break;
|
|
|
|
}
|
2008-03-03 22:01:56 +00:00
|
|
|
GST_DEBUG_OBJECT (best->collect.pad, "best pad - buffer ts %"
|
|
|
|
GST_TIME_FORMAT " dur %" GST_TIME_FORMAT,
|
|
|
|
GST_TIME_ARGS (GST_BUFFER_TIMESTAMP (best->buffer)),
|
|
|
|
GST_TIME_ARGS (GST_BUFFER_DURATION (best->buffer)));
|
2006-01-23 10:44:03 +00:00
|
|
|
|
2008-01-08 19:57:23 +00:00
|
|
|
/* make note of first and last encountered timestamps, so we can calculate
|
|
|
|
* the actual duration later when we send an updated header on eos */
|
2008-03-03 22:01:56 +00:00
|
|
|
if (GST_BUFFER_TIMESTAMP_IS_VALID (best->buffer)) {
|
|
|
|
GstClockTime start_ts = GST_BUFFER_TIMESTAMP (best->buffer);
|
|
|
|
GstClockTime end_ts = start_ts;
|
|
|
|
|
|
|
|
if (GST_BUFFER_DURATION_IS_VALID (best->buffer))
|
|
|
|
end_ts += GST_BUFFER_DURATION (best->buffer);
|
|
|
|
else if (best->track->default_duration)
|
|
|
|
end_ts += best->track->default_duration;
|
|
|
|
|
2008-05-08 08:15:34 +00:00
|
|
|
if (!GST_CLOCK_TIME_IS_VALID (best->end_ts) || end_ts > best->end_ts)
|
2008-03-03 22:01:56 +00:00
|
|
|
best->end_ts = end_ts;
|
|
|
|
|
|
|
|
if (G_UNLIKELY (best->start_ts == GST_CLOCK_TIME_NONE ||
|
|
|
|
start_ts < best->start_ts))
|
|
|
|
best->start_ts = start_ts;
|
2008-01-08 19:57:23 +00:00
|
|
|
}
|
|
|
|
|
2006-01-23 10:44:03 +00:00
|
|
|
/* write one buffer */
|
|
|
|
ret = gst_matroska_mux_write_data (mux, best);
|
|
|
|
} while (ret == GST_FLOW_OK && !popped);
|
|
|
|
|
|
|
|
return ret;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* gst_matroska_mux_change_state:
|
|
|
|
* @element: #GstMatroskaMux
|
|
|
|
* @transition: State change transition.
|
|
|
|
*
|
|
|
|
* Change the muxer state.
|
|
|
|
*
|
|
|
|
* Returns: #GstStateChangeReturn
|
|
|
|
*/
|
2005-09-02 15:44:50 +00:00
|
|
|
static GstStateChangeReturn
|
|
|
|
gst_matroska_mux_change_state (GstElement * element, GstStateChange transition)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
2005-10-14 12:43:30 +00:00
|
|
|
GstStateChangeReturn ret;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
GstMatroskaMux *mux = GST_MATROSKA_MUX (element);
|
|
|
|
|
2005-09-02 15:44:50 +00:00
|
|
|
switch (transition) {
|
2005-10-14 12:43:30 +00:00
|
|
|
case GST_STATE_CHANGE_NULL_TO_READY:
|
|
|
|
break;
|
|
|
|
case GST_STATE_CHANGE_READY_TO_PAUSED:
|
Fix for collect pads API change. Also fix textoverlay state change function.
Original commit message from CVS:
* ext/cairo/gsttextoverlay.c: (gst_text_overlay_finalize),
(gst_text_overlay_init), (gst_text_overlay_text_pad_linked),
(gst_text_overlay_text_pad_unlinked), (gst_text_overlay_pop_video),
(gst_text_overlay_pop_text), (gst_text_overlay_collected),
(gst_text_overlay_change_state):
* gst/matroska/matroska-mux.c: (gst_matroska_mux_init),
(gst_matroska_mux_reset), (gst_matroska_mux_request_new_pad),
(gst_matroska_mux_best_pad), (gst_matroska_mux_change_state):
* gst/smpte/gstsmpte.c: (gst_smpte_init), (gst_smpte_collected):
* gst/videomixer/videomixer.c: (gst_videomixer_init),
(gst_videomixer_request_new_pad), (gst_videomixer_fill_queues),
(gst_videomixer_change_state):
Fix for collect pads API change. Also fix textoverlay state
change function.
2005-11-21 12:13:48 +00:00
|
|
|
gst_collect_pads_start (mux->collect);
|
2005-10-14 12:43:30 +00:00
|
|
|
break;
|
|
|
|
case GST_STATE_CHANGE_PAUSED_TO_PLAYING:
|
|
|
|
break;
|
2006-05-23 20:15:04 +00:00
|
|
|
case GST_STATE_CHANGE_PAUSED_TO_READY:
|
|
|
|
gst_collect_pads_stop (mux->collect);
|
|
|
|
break;
|
2005-10-14 12:43:30 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = GST_ELEMENT_CLASS (parent_class)->change_state (element, transition);
|
|
|
|
|
|
|
|
switch (transition) {
|
|
|
|
case GST_STATE_CHANGE_PLAYING_TO_PAUSED:
|
|
|
|
break;
|
2005-09-02 15:44:50 +00:00
|
|
|
case GST_STATE_CHANGE_PAUSED_TO_READY:
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
gst_matroska_mux_reset (GST_ELEMENT (mux));
|
|
|
|
break;
|
2005-10-14 12:43:30 +00:00
|
|
|
case GST_STATE_CHANGE_READY_TO_NULL:
|
|
|
|
break;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2005-10-14 12:43:30 +00:00
|
|
|
return ret;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_set_property (GObject * object,
|
|
|
|
guint prop_id, const GValue * value, GParamSpec * pspec)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstMatroskaMux *mux;
|
|
|
|
|
|
|
|
g_return_if_fail (GST_IS_MATROSKA_MUX (object));
|
|
|
|
mux = GST_MATROSKA_MUX (object);
|
|
|
|
|
|
|
|
switch (prop_id) {
|
2005-10-14 12:43:30 +00:00
|
|
|
case ARG_WRITING_APP:
|
|
|
|
if (!g_value_get_string (value)) {
|
|
|
|
GST_WARNING_OBJECT (mux, "writing-app property can not be NULL");
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
g_free (mux->writing_app);
|
gst/matroska/ebml-write.c: Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
Original commit message from CVS:
* gst/matroska/ebml-write.c: (gst_ebml_write_float):
Use GDOUBLE_TO_BE() instead of (probably slower) custom code.
* gst/matroska/matroska-demux.c: (gst_matroska_demux_base_init),
(gst_matroska_demux_class_init), (gst_matroska_demux_init),
(gst_matroska_track_free), (gst_matroska_demux_encoding_cmp),
(gst_matroska_demux_read_track_encodings),
(gst_matroska_demux_add_stream),
(gst_matroska_demux_handle_src_query),
(gst_matroska_demux_init_stream),
(gst_matroska_demux_parse_index_cuetrack),
(gst_matroska_demux_parse_index_pointentry),
(gst_matroska_demux_parse_info),
(gst_matroska_demux_parse_metadata_id_simple_tag),
(gst_matroska_demux_parse_metadata),
(gst_matroska_demux_add_wvpk_header), (gst_matroska_decode_buffer),
(gst_matroska_demux_parse_blockgroup_or_simpleblock),
(gst_matroska_demux_parse_cluster),
(gst_matroska_demux_parse_contents_seekentry),
(gst_matroska_demux_loop_stream_parse_id),
(gst_matroska_demux_loop), (gst_matroska_demux_video_caps),
(gst_matroska_demux_audio_caps),
(gst_matroska_demux_subtitle_caps):
* gst/matroska/matroska-demux.h:
* gst/matroska/matroska-ids.c:
(gst_matroska_track_init_subtitle_context):
* gst/matroska/matroska-ids.h:
* gst/matroska/matroska-mux.c: (gst_matroska_mux_base_init),
(gst_matroska_mux_class_init), (gst_matroska_mux_init),
(gst_matroska_mux_create_uid), (gst_matroska_mux_reset),
(gst_matroska_mux_video_pad_setcaps),
(gst_matroska_mux_audio_pad_setcaps),
(gst_matroska_mux_subtitle_pad_setcaps),
(gst_matroska_mux_request_new_pad),
(gst_matroska_mux_track_header), (gst_matroska_mux_start),
(gst_matroska_mux_write_simple_tag), (gst_matroska_mux_finish),
(gst_matroska_mux_write_data), (gst_matroska_mux_collected),
(gst_matroska_mux_set_property):
Add many FIXMEs/TODOs all over the matroska muxer and demuxer
elements, do some checks for valid values in the demuxer, handle
tracktimecodescale in the demuxer, set correct default values for all
settings in the demuxer, review and add all missing matroska
IDs and some more raw YUV formats, and some trivial cleanup.
2008-06-10 10:44:53 +00:00
|
|
|
mux->writing_app = g_value_dup_string (value);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
break;
|
2005-10-28 15:32:48 +00:00
|
|
|
case ARG_MATROSKA_VERSION:
|
|
|
|
mux->matroska_version = g_value_get_int (value);
|
|
|
|
break;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
default:
|
|
|
|
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_get_property (GObject * object,
|
|
|
|
guint prop_id, GValue * value, GParamSpec * pspec)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
GstMatroskaMux *mux;
|
|
|
|
|
|
|
|
g_return_if_fail (GST_IS_MATROSKA_MUX (object));
|
|
|
|
mux = GST_MATROSKA_MUX (object);
|
|
|
|
|
|
|
|
switch (prop_id) {
|
2005-10-14 12:43:30 +00:00
|
|
|
case ARG_WRITING_APP:
|
|
|
|
g_value_set_string (value, mux->writing_app);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
break;
|
2005-10-28 15:32:48 +00:00
|
|
|
case ARG_MATROSKA_VERSION:
|
|
|
|
g_value_set_int (value, mux->matroska_version);
|
|
|
|
break;
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
default:
|
|
|
|
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
gboolean
|
2004-03-14 22:34:33 +00:00
|
|
|
gst_matroska_mux_plugin_init (GstPlugin * plugin)
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
{
|
|
|
|
return gst_element_register (plugin, "matroskamux",
|
2009-05-07 17:10:08 +00:00
|
|
|
GST_RANK_PRIMARY, GST_TYPE_MATROSKA_MUX);
|
Riff, EBML, fourcc etc. work. Not fully finished, but better than what we used to have and definately worth a first b...
Original commit message from CVS:
Riff, EBML, fourcc etc. work. Not fully finished, but better than
what we used to have and definately worth a first broad testing.
I've revived rifflib. Rifflib used to be a bytestream-for-riff, which
just dup'ed bytestream. I've rewritten rifflib to be a modern riff-
chunk parser that uses bytestream fully, plus adds some extra functions
so that riff file parsing becomes extremely easy. It also contains some
small usability functions for strh/strf and metadata parsing. Note that
it doesn't use the new tagging yet, that's a TODO.
Avidemux has been rewritten to use this. I think we all agreed that
avidemux was pretty much a big mess, which is because it used all
sort of bytestream magic all around the place. It was just ugly.
This is a lot nicer, very complete and safe. I think this is far more
robust than what the old avidemux could ever have been. Of course, it
might contain bugs, please let me know.
EBML writing has also been implemented. This is useful for matroska.
I'm intending to modify avidemux (with a riffwriter) similarly. Maybe
I'll change wavparse/-enc too to use rifflib.
Lastly, several plugins have been modified to use rifflib's fourcc
parsing instead of their own. this puts fourcc parsing in one central
place, which should make it a lot simpler to add new fourccs. We might
want to move this to its own lib instead of rifflib.
Enjoy!
2003-12-07 20:00:41 +00:00
|
|
|
}
|