* commit '2d7d91f06d6a1d243dc74c96d3389ee237a3b906':
svq1enc: Set picture_structure correctly
mpegvideo: remove last_picture_ptr / h264 assert.
mpegvideo_enc: remove assert that has become obsolete with the new API
elbg: Fix an assert
* qatar/master:
configure: add support for bdver1 and bdver2 CPU types.
avio: make avio_close NULL the freed buffer
pixdesc: cosmetics
proresenc: Don't free a buffer not owned by the codec
proresenc: Write the full value in one put_bits call
adpcmenc: Calculate the IMA_QT predictor without overflow
x86: Add convenience macros to check for CPU extensions and flags
x86: h264dsp: drop some unnecessary ifdefs around prototype declarations
mss12: merge decode_pixel() and decode_top_left_pixel()
mss12: reduce SliceContext size from 1067 to 164 KB
mss12: move SliceContexts out of the common context into the codec contexts
Martin Storsjö [Tue, 4 Sep 2012 11:57:45 +0000 (14:57 +0300)]
proresenc: Don't free a buffer not owned by the codec
The data in coded_frame isn't allocated using get_buffer, but
is copied from the input frame to the encoder, so we should
not try to free it ourselves.
This fixes an assert failure when running in debug mode.
Martin Storsjö [Tue, 4 Sep 2012 11:31:52 +0000 (14:31 +0300)]
proresenc: Write the full value in one put_bits call
Previously, the put_bits call writing the value wrote a value
larger than the number of bits specified, failing asserts
in debug mode. There was no actual bitstream writer corruption,
since the overwritten bit already always was set to 1.
adpcmenc: Calculate the IMA_QT predictor without overflow
Previously, the value given to put_bits was 10 bits long for positive
predictors, even though 9 bits were to be written. The extra bit could
in some cases overwrite existing bits in the bitstream writer cache.
This fixes a failed assert in put_bits.h, when running a version
built with -DDEBUG.
The fate test result gets slightly improved, thanks to getting rid
of the overwritten bits in the bitstream writer cache.
While not that usefull, we can as well keep it until it breaks.
When it breaks for whatever reason ill likely remove it
Sorry for the revert spam, i had not realized this code compiles
and works fine.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master:
build: avcodec: order cosmetics
intrax8dsp: Add missing #include to make header standalone
Delete motion-test tool; it is of doubtful utility.
configure: Treat warnings as errors in check_yasm()
yasm generates a warning instead of an error when "CPU amdnop" is not supported, so the check always succeeds.
HAVE_CPUNOP as true when the available yasm version doesn't support amdnop doesn't make sense.
Check http://fate.ffmpeg.org/log.cgi?time=20120901143315&log=compile&slot=x86_64-archlinux-gcc-oldyasm to see the amount of warnings it generates.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
fix flv new stream handling after head already parsed
This fixes stream lookup in flv demuxer. When used with librtmp
protocol streams are sometimes added after head arrived. If the
first stream added in flv header reader is Audio stream then it
is messed with Video stream added later in the code patched.
The result is I have 2 Audio streams (first of them is said to have
a video codec like h264) instead of Audio/Video pair.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* cus/stable:
ffplay: only free vfilters on exit
ffplay: replace SDL_delay in read thread with SDL_CondWait
ffplay: requeue last picture on forced video refresh
libavfilter: Support the forks ABI for buffer sinks
With this change avconv compiled against libav and linked to ffmpegs libs
will run through the whole fate testsuite without any crashes.
857 tests pass, the remaining tests fail one way or another, which is
to be expected as avconv is not a drop in replacement for ffmpeg
The testsuite used was the ffmpeg fate testsuite, not libavs.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Philip Langdale [Wed, 23 May 2012 04:50:02 +0000 (21:50 -0700)]
matroskadec, matroskadec, srtenc: Read/Write duration for subtitles.
After much discussion and back-and-forth, we reached the conclusion
that matroska uses convergence_duration for subtitle duration because
a 32bit value isn't large enough to store the duration if sub-micro-second
timebases are used. Matroska may not be the only one that supports these
timebases, but it's certainly the only one that ffmpeg attempts to support
in this way.
The long term solution that we seemed to reach was that if we encounter
a matroska file with a sub-micro-second timebase, we should internally
scale it up to at least micro-second, and then duration can be used
normally. This suggests that on the encode side, we should not allow
generation of files with sub-micro-second timebases, but that's a separate
issue.
That being a non-trivial change, and the subtitle interoperability breakage
being very real, I'm re-submitting this small change for consideration.
In this diff, we make sure that duration is populated by the matroska
demuxer, and that convergence_duration is respected in matroskaenc and
srtenc, but that duration is used otherwise. This ends up being a strict
improvement - pipelines that use convergence duration are unchanged, and
ones that are currently broken due to the duration mismatch will start
working - except for the ones with the extreme timebases, but those were
already broken.
Signed-off-by: Philip Langdale <philipl@overt.org>
Marton Balint [Wed, 29 Aug 2012 21:56:33 +0000 (23:56 +0200)]
ffplay: only free vfilters on exit
Freeing it in the end of the video thread is not a good idea, because we still
may need the filter names for the next video thread, in order to apply the
filters after chaning the video stream.
Marton Balint [Mon, 27 Aug 2012 22:00:15 +0000 (00:00 +0200)]
ffplay: replace SDL_delay in read thread with SDL_CondWait
When the audio queue was empty, it was not filled until the 10ms delay expired
in the read thread. This patch changes the delay method with a condition wait,
which reacts to an empty queue a lot faster, therefore the audio buffer
underruns become less common especially after seeking.
projects / ffmpeg / log