Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video):
Use gst_pad_alloc_buffer_and_set_caps() to make sure we get
a buffer with caps that we can work with (i.e. the pad's caps).
Add non-keyframe video frames to the index too but without the
keyframe flag.
Add audio frames to the index only if we have no video stream.
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video):
Create pads from the pad templates, use fixed caps on them
and only activate them after the caps are set.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_loop):
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_timestamp):
* gst/flv/gstflvparse.h:
Get an approximate duration of the file by looking at the timestamp
of the last tag in pull mode. If we get (maybe better) duration from
metadata later we'll use that instead.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_pull_range),
(gst_flv_demux_pull_tag), (gst_flv_demux_pull_header):
Refactor _pull_range() logic with checks into a seperate function
to make things a bit more readable.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_chain),
(gst_flv_demux_base_init):
Use gst_element_class_set_details_simple().
If we get GST_FLOW_NOT_LINKED in the parse loop but at least
one of the pads is linked continue the loop.
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_audio_negotiate),
(gst_flv_parse_tag_audio), (gst_flv_parse_video_negotiate):
Correct caps for video codec id 5: It's On2 VP6 with alpha channel
which needs a different decoder and has different caps.
Add support for audio codec id 14, which is MP3 with 8kHz sampling
rate.
Fix endianness and signedness for raw audio codec ids.
Add support for alaw and mulaw audio.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_chain):
Go out of the parse loop as soon as we get an error instead
of parsing until the GstAdapter is empty.
Add some explanations about the header and tag size.
Don't print synchronizing message if everything is fine.
Original commit message from CVS:
* gst/flv/Makefile.am:
* gst/flv/gstflvdemux.c: (plugin_init):
* gst/flv/gstflvmux.c: (gst_flv_mux_base_init),
(gst_flv_mux_class_init), (gst_flv_mux_init),
(gst_flv_mux_finalize), (gst_flv_mux_reset),
(gst_flv_mux_handle_src_event), (gst_flv_mux_handle_sink_event),
(gst_flv_mux_video_pad_setcaps), (gst_flv_mux_audio_pad_setcaps),
(gst_flv_mux_request_new_pad), (gst_flv_mux_release_pad),
(gst_flv_mux_write_header), (gst_flv_mux_write_buffer),
(gst_flv_mux_collected), (gst_flv_mux_change_state):
* gst/flv/gstflvmux.h:
Add first version of a FLV muxer. The only missing feature is writing
of stream metadata.
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_metadata_item),
(gst_flv_parse_tag_script):
Handle NULL returns from FLV_GET_STRING() more gracefully. Fixes
crash caused by a strlen on a NULL string (#527622).
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video):
Don't strdup (and thus leak) codec name strings when passing
them to gst_tag_list_add().
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video):
Fix list of supported and known codecs.
Emit tag with the codec name so it gets properly reported in totem and
other applications.
Original commit message from CVS:
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video):
Output segment with proper 'stop' value, makes flvdemux 100% compatible
with gnonlin.
Original commit message from CVS:
* gst/flv/gstflvparse.c:
Add mapping for Nellymoser ASAO audio codec.
(gst_flv_parse_tag_audio), (gst_flv_parse_tag_video): Make sure we
actually have data to read at the end of the tag. This avoids trying
to allocate negative buffers.
Original commit message from CVS:
2007-10-22 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video), (gst_flv_parse_tag_type): Don't
emit no-more-pads for single pad scenarios as the header
is definitely not reliable. We emit them for 2 pads scenarios
though to speed up media discovery.
Original commit message from CVS:
2007-09-27 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): I got it wrong again, audio rate
was not detected correctly in all cases.
Original commit message from CVS:
2007-09-26 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): codec_data is needed for every tag
not just the first one. (Fix a stupid bug i introduced without
testing)
Original commit message from CVS:
2007-09-26 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvparse.c: (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): Fix bit masks operations to be
sure we detect the codec_tags and sample rates correctly.
Fix raw audio caps generation.
Original commit message from CVS:
2007-08-27 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvdemux.c: (gst_flv_demux_handle_seek_pull):
Make sure we initialize the seek result.
Original commit message from CVS:
2007-08-22 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvdemux.c: (gst_flv_demux_cleanup),
(gst_flv_demux_pull_tag):
* gst/flv/gstflvdemux.h:
* gst/flv/gstflvparse.c: (gst_flv_parse_metadata_item),
(gst_flv_parse_tag_script), (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): Handle pixel aspect ratio through
metadata tags like ASF does. Fluendo muxer supports this and
Flash players can support it as well this way.
Original commit message from CVS:
2007-08-22 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvdemux.c: (gst_flv_demux_pull_tag):
* gst/flv/gstflvparse.c: (gst_flv_parse_metadata_item),
(gst_flv_parse_tag_script), (gst_flv_parse_tag_audio),
(gst_flv_parse_tag_video): Make sure we don't try filling up the
index if no times object was parsed. Fix the way we decide to
push
tags and emit no-more-pads. Fix some printf typing in debugging.
Original commit message from CVS:
* gst/flv/gstflvdemux.c: (gst_flv_demux_set_index),
(gst_flv_demux_get_index):
Fix locking and refcounting on the index.
Original commit message from CVS:
2007-08-14 Julien MOUTTE <julien@moutte.net>
* gst/flv/gstflvdemux.c: (gst_flv_demux_cleanup),
(gst_flv_demux_adapter_flush), (gst_flv_demux_chain),
(gst_flv_demux_pull_tag), (gst_flv_demux_do_seek),
(gst_flv_demux_handle_seek), (gst_flv_demux_sink_event),
(gst_flv_demux_src_event), (gst_flv_demux_query),
(gst_flv_demux_change_state), (gst_flv_demux_set_index),
(gst_flv_demux_get_index), (gst_flv_demux_dispose),
(gst_flv_demux_class_init): First method for seeking in pull
mode using the index built step by step or coming from metadata.
* gst/flv/gstflvdemux.h:
* gst/flv/gstflvparse.c: (FLV_GET_STRING),
(gst_flv_parse_metadata_item), (gst_flv_parse_tag_script),
(gst_flv_parse_tag_audio), (gst_flv_parse_tag_video): Parse
more metadata types and keyframes index.
Original commit message from CVS:
2007-07-19 Julien MOUTTE <julien@moutte.net>
* configure.ac:
* gst/flv/Makefile.am:
* gst/flv/gstflvdemux.c: (gst_flv_demux_flush),
(gst_flv_demux_cleanup), (gst_flv_demux_chain),
(gst_flv_demux_pull_tag), (gst_flv_demux_pull_header),
(gst_flv_demux_seek_to_prev_keyframe), (gst_flv_demux_loop),
(gst_flv_demux_sink_activate),
(gst_flv_demux_sink_activate_push),
(gst_flv_demux_sink_activate_pull), (gst_flv_demux_sink_event),
(gst_flv_demux_change_state), (gst_flv_demux_dispose),
(gst_flv_demux_base_init), (gst_flv_demux_class_init),
(gst_flv_demux_init), (plugin_init):
* gst/flv/gstflvdemux.h:
* gst/flv/gstflvparse.c: (FLV_GET_BEUI24), (FLV_GET_STRING),
(gst_flv_demux_query_types), (gst_flv_demux_query),
(gst_flv_parse_metadata_item), (gst_flv_parse_tag_script),
(gst_flv_parse_tag_audio), (gst_flv_parse_tag_video),
(gst_flv_parse_tag_type), (gst_flv_parse_header):
* gst/flv/gstflvparse.h: Adds a first draft of an FLV demuxer.
It does not do seeking yet, it supports pull and push mode so
YES
you can use it to play youtube videos directly from an HTTP uri.
Not so much testing done yet but it parses metadata, reply to
duration queries, etc...
The test creates buffers with non-silence, sets the GAP
flag on it and expects rganalysis to ignore the content and assume silence.
That's not the way how GAP buffers should be used, if the GAP flag is set
elements *can* assume that they only contain silence but they're not *required*
to assume that. The GAP flag must only be set on silence buffers.
Fixes bug #582252.
First of all a keyframe seek should be done to the
keyframe right before the requested position and not
to the keyframe that is nearest to the requested position.
Use per track index arrays and use our new binary search function
from core to speed up the search.
It was previously sending the bogus buffer which was returned from
the bufferalloc (required for reverse negotiation apparently) instead
of the pending buffer.
