For audio always add the minimum ffmpeg buffer size, for video
use the same weird buffer size as they use in ffmpeg.c:
width*height*6 + 200
Also make setting of the buffer-size property a no-op.
Fixes bug #593651.
PixFmt that are declared in AVCodec.pix_fmts are ones which are 'officially'
declared as being supported. We should therefore not have to create a
AVCodecContext and open an encoder to know if it's supported or not.
Also, doing it this way allows us to better pickup configuration overrides
we have in gstffmpegcodecmap for some codecs (like restrictions on width,
height, framerate like it's the case for dnxhd).
Fixes#575545
If the same instance of the plugin is asked to be initialised more that once,
instances after the first one do not register the elements properly and the
elements become not usable.
For example, if you call gst_update_registry (), is not possible to create
elements after the call since the plugin is asked to be initialised again and
does not register the elements.
Fixes#584291
Rewrite the audio encoders to use the right API functions of ffmpeg. Also get
rid of the handrolled cache and use adapter instead for formats that require
fixed frame_size as input.
For a given AVCodec, when the sample_fmts field is non-NULL, that means that
that codec can only handle a specific set of SampleFormat.
With this patch, we now look for its presence and create the proper pad template
caps.
Fixes#569441
Original commit message from CVS:
Based on a patch by: Alexis Ballier <aballier at gentoo dot org>
* configure.ac:
* ext/ffmpeg/gstffmpeg.c:
* ext/ffmpeg/gstffmpeg.h:
* ext/ffmpeg/gstffmpegaudioresample.c:
* ext/ffmpeg/gstffmpegcfg.c: (gst_ffmpeg_flags_get_type),
(gst_ffmpeg_cfg_init):
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_to_pixfmt), (gst_ffmpeg_caps_with_codecid):
* ext/ffmpeg/gstffmpegcodecmap.h:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_setcaps),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdeinterlace.c:
* ext/ffmpeg/gstffmpegdemux.c:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps),
(gst_ffmpegenc_setcaps), (gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegmux.c:
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_seek):
* ext/libpostproc/gstpostproc.c:
* ffmpegrev:
Update ffmpeg/swscale snapshot to the latest revision and adjust
to API changes. Fixes bug #556405.
Require libavutil for swscale too when building with an external
ffmpeg and fix includes for external ffmpeg.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ff_vid_caps_new),
(gst_ff_aud_caps_new), (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_codectype_to_caps):
* ext/ffmpeg/gstffmpegcodecmap.h:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_negotiate),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps),
(gst_ffmpegenc_register):
Add some more width/height/channels/rate limitations to caps
to cater for more automagic negotiation. Addresses #532422.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcfg.c: (gst_ffmpeg_cfg_fill_context):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
Latest ffmpeg revision's avcodec_close frees more; use safer coding to
prevent double free and other related segfaults.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_caps_to_pixfmt):
Recognize video/x-raw-gray and map to proper pixfmt.
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps),
(gst_ffmpegenc_chain_video):
Fail negotiation if pixfmt cannot be determined from input caps.
Prevent segfault accessing non-existant coded_frame, provide some
warning debug output instead.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ff_vid_caps_new),
(gst_ffmpeg_pixfmt_to_caps), (gst_ffmpeg_caps_with_codecid):
Add more specific width/height for DV video so that it negotiates more
automatically.
Try to avoid accessing NULL caps.
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
Add some more debuggin to the caps generation.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_base_init),
(gst_ffmpegdec_setcaps), (gst_ffmpegdec_video_frame):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init):
When doing QoS, don't drop the frame before decoding because we might
drop an important reference frame, just make the decoder hurry_up on
this frame.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_chain_video),
(gst_ffmpegenc_flush_buffers):
Fix the encoders by actually copying the encoded output data from the
working area into the output buffer.
Fixes: #453135.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (ffmpegenc_setup_working_buf),
(gst_ffmpegenc_chain_video), (gst_ffmpegenc_flush_buffers),
(gst_ffmpegenc_change_state):
* ext/ffmpeg/gstffmpegenc.h:
Instead of allocating 512KB buffers and then setting BUFFER_SIZE
to what ffmpeg produced, we're better off by far to use a single
working buffer and copy data out into neat little buffers.
This prevents exhorbitant virtual memory wastage in the form
of allocated but untouched buffers.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_base_init),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_base_init),
(gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init),
(gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_base_init),
(gst_ffmpegmux_register):
Don't use GINT_TO_POINTER and GPOINTER_TO_INT with GTypes, this
can break horribly if sizeof(GType) happens to be bigger than
sizeof(int), because GPOINTER_TO_INT might then chop off some bits
of our GType (the reason this seems to works nevertheless is the
put-current-type-also-as-value-0-into-the-hash-table-as-fallback
hack used in the current code).
In any case, instead of just fixing this, let's not use a hash table
with GTypes here at all. g_type_{set|get}_qdata() seems to do the
job just as well.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_me_method_get_type):
Make the name of the motion-estimation method enum values meaningful.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_class_init),
(gst_ffmpegenc_finalize):
Turn dispose function into finalize function to avoid double-frees
and potential crashes caused by them in some circumstances.
Fixes#392395.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_register):
My console is not your TODO list.
If it were your TODO list, you would have DONE it by now.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps),
(gst_ffmpegenc_setcaps), (gst_ffmpegenc_register):
MPEG4 standard does not accept framerates with a denominator greater
than (1<<16 - 1) (according to the ffmpeg encoder...).
We therefore scale down the framerate.
Small indentation fixes using gst-indent.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcfg.c:
FLV is in fact h263 and can take the same parameters as other mpeg
derivatives.
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_with_codecid):
Add pixel format to video/x-dv,systemstream=False
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_class_init),
(gst_ffmpegenc_init):
Use EPZS (Enhanced Predictive Zonal Search) as the default motion
estimation method, since it's the best quality to speed compromise.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c:
(gst_ffmpeg_formatid_get_codecids):
Added some codec mappings for dv and mov muxers.
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps),
(gst_ffmpegenc_register):
Don't attempt to add caps to a NULL caps.
Add video/x-raw-gray to the input raw video types.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_init):
Properly initialize ->lmin, ->lmax and ->max_key_interval so we don't
end up passing completely bogus data to the AVCodecContext.
Fixes#355584
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_get_stream):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps):
Printf format fixes.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_register):
Don't leak caps in some cases and make build bots happy again.
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet dot be>
* ext/ffmpeg/Makefile.am:
* ext/ffmpeg/gstffmpegcfg.c: (gst_ffmpeg_pass_get_type),
(gst_ffmpeg_lim_pass_get_type), (gst_ffmpeg_mb_decision_get_type),
(gst_ffmpeg_mb_cmp_get_type), (gst_ffmpeg_dct_algo_get_type),
(gst_ffmpeg_idct_algo_get_type), (gst_ffmpeg_quant_type_get_type),
(gst_ffmpeg_pre_me_get_type), (gst_ffmpeg_pred_method_get_type),
(gst_ffmpeg_flags_get_type), (gst_ffmpeg_cfg_init),
(gst_ffmpeg_cfg_codec_has_pspec),
(gst_ffmpeg_cfg_install_property), (gst_ffmpeg_cfg_set_property),
(gst_ffmpeg_cfg_get_property), (gst_ffmpeg_cfg_set_defaults),
(gst_ffmpeg_cfg_fill_context):
* ext/ffmpeg/gstffmpegcfg.h:
Added helper object to handle extra properties.
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_class_init),
(gst_ffmpegenc_init), (gst_ffmpegenc_dispose),
(gst_ffmpegenc_getcaps), (gst_ffmpegenc_setcaps),
(gst_ffmpegenc_chain_video), (gst_ffmpegenc_chain_audio),
(gst_ffmpegenc_flush_buffers), (gst_ffmpegenc_event_video),
(gst_ffmpegenc_set_property), (gst_ffmpegenc_get_property),
(gst_ffmpegenc_change_state), (gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegenc.h:
Exposes (quite some of) the AVCodecContext configurable
settings, as well as some additional (though simple) logic for b-frame
delay handling.
Most of the properties are only enabled for mpeg4-like codecs.
Fixes#344583.
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet dot be>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps):
If we are not operating in RTP mode, leave the default rtp encoder
settings as they are so that we don't unintentionally activate it.
Fixes#351415.
Original commit message from CVS:
Reviewed by : Edward Hervey <edward@fluendo.com>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_chain_audio):
Fix segmentation fault and memleak.
Fixes#332995 and #333001
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_caps_to_pixfmt):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
respect pixel-aspect-ratio when encoding.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_class_init),
(gst_ffmpegenc_init), (gst_ffmpegenc_setcaps):
debug and magic number fixes
Original commit message from CVS:
Reviewed by : Edward Hervey <edward@fluendo.com>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_chain_audio):
Caps weren't set on outgoing audio buffers.
Closes#327861
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps):
we have better ways of debugging caps
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps):
ye gods. get_template_caps doesn't return a ref, so when the
pad was not linked, crashes would ensue. Serves me right.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps):
Do not limit quantizer values so we can get realy low
bitrates.
Original commit message from CVS:
* ext/ffmpeg/gstffmpeg.c: (gst_ffmpeg_log_callback), (plugin_init):
* ext/ffmpeg/gstffmpeg.h:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
shut up the logging while it's probing for formats
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps),
(gst_ffmpegenc_chain_video), (gst_ffmpegenc_change_state):
Fix return values from the setcaps function.
Original commit message from CVS:
Reviewed by: Luca Ognibene <luogni@tin.it>
* ext/ffmpeg/gstffmpegenc.c (gst_ffmpegenc_chain_video):
set the GST_BUFFER_DELTA_UNIT while encoding
fix#308804
Original commit message from CVS:
* configure.ac:
scale requires -libs (video.h).
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
Don't crash if open failed.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c (gst_ffmpegenc_chain_video):
* ext/ffmpeg/gstffmpegdec.c (gst_ffmpegdec_chain):
Makes ffmpegdec and ffmpegenc use these functions
and remove a comment from a previuos patch
* ext/ffmpeg/gstffmpegcodecmap.h:
Add two utility functions to convert the pts from/to
the ffmpeg format
Original commit message from CVS:
Reviewed by: Ronald S. Bultje <rbultje@ronald.bitfreak.net>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_getcaps):
Don't crash if avcodec_open failed (#301717).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_with_codecid), (gst_ffmpeg_caps_to_codecid),
(gst_ffmpeg_get_codecid_longname):
Add ALAC, shorten and some more in new FFMPEG snapshot of
21/4, somewhere this afternoon.
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_register):
Add warnings to decoder registration to easier add new ones
during snapshot updates.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_to_codecid):
Add AMR mimetypes (since those are recognized in ffmpeg).
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_frame),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_register):
Register law codecs too, add time debug. Use ANY for template caps
since we don't necessarily have available decoders for recognized
formats.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_init),
(gst_ffmpegenc_link), (gst_ffmpegenc_chain_video):
Remove EMU_EDGE, since that really fixes#162905. Revert
previous hacks.
Original commit message from CVS:
Reviewed by: Ronald S. Bultje <rbultje@ronald.bitfreak.net>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_link),
(gst_ffmpegenc_chain_video):
Workaround memory issues (#162905).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init),
(gst_ffmpegenc_init), (gst_ffmpegenc_getcaps),
(gst_ffmpegenc_link):
Move setup of sink caps (accepted formats) to a separate function
rather than doing it on initialization.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_init),
(gst_ffmpegenc_chain_video):
Set chain function before adding pad, fix memleak on error, error
out on encode errors because none of subsequent frames will work
any better (this is not input dependent).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c
* ext/ffmpeg/gstffmpegenc.c
change plugins category to meet our new naming standard :
Codec/Encoder|Decoder/Audio|Video
suggested by bilboed, approved by Company
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_get_palette),
(gst_ffmpeg_set_palette), (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_pixfmt_to_caps), (gst_ffmpeg_caps_to_smpfmt),
(gst_ffmpeg_caps_to_pixfmt), (gst_ffmpeg_caps_with_codectype),
(gst_ffmpeg_caps_with_codecid), (gst_ffmpeg_caps_to_codecid):
* ext/ffmpeg/gstffmpegcodecmap.h:
Change some function names to reflect that they don't really
_return_ something, but rather _use_ something to fill a
AVCodecContext. s/to/with/. Restructure the extradata handling,
it's now not picking up the type from the caps but rather
using the type as provided in the function. This is a lot
cleaner. Implement MS RLE palette pickup.
* ext/ffmpeg/gstffmpegcolorspace.c: (gst_ffmpegcsp_pad_link):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_connect):
Sync with the above function name changes.
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_close),
(gst_ffmpegdec_open), (gst_ffmpegdec_connect),
(gst_ffmpegdec_chain), (gst_ffmpegdec_change_state):
Add some hacks to convert palette-based raw image formats to
RGBA32. Ugly, but I don't know how else to handle palette-based
RGB, since img_convert() (and thus ffcolorspace) doesn't accept
a palette as argument.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_base_init):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init):
Fix category.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_class_init),
(gst_ffmpegdemux_init), (gst_ffmpegdemux_close):
Slightly rewrite the init function, old one was ugly. Also remove
dispose function since it's obsolete, the change_state function
cleans up already. Also fixes segfault in test-app
gstreamer/tests/instantiate/create
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps):
* ext/ffmpeg/gstffmpegcodecmap.h:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_chain),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_add),
(gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_connect),
(gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_register):
Add encode property to codec mapper, so we can get better-fitting
caps. Also use this everywhere. ;). Lastly, add some new codecIDs
to our list of recognized ones.
Original commit message from CVS:
2004-01-16 Ronald Bultje <rbultje@ronald.bitfreak.net>
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_connect):
Link with right caps (else, it segfaults).
* ext/mplex/gstmplexjob.cc:
Fix for slight API change in 1.6.1.93 release of mjpegtools.
