Anton Khirnov [Tue, 24 Jun 2014 05:39:35 +0000 (07:39 +0200)]
dump_stream: print the timebase as is, do not reduce it
It makes more sense to print the timebase exactly as it is set. Also,
this avoids a divide by zero when av_dump_format() is called on a format
context before writing the header.
Diego Biurrun [Wed, 25 Jun 2014 11:35:30 +0000 (04:35 -0700)]
x86: h264dsp: Fix link failure with optimizations disabled
With optimzations disabled compilers have trouble doing dead code
elimination on 'if (foo && 0)' expressions, while 'if (0 && foo)'
still works, so use the latter to avoid problems.
Anton Khirnov [Sun, 18 May 2014 10:12:59 +0000 (12:12 +0200)]
lavf: switch to AVStream.time_base as the hint for the muxer timebase
Previously, AVStream.codec.time_base was used for that purpose, which
was quite confusing for the callers. This change also opens the path for
removing AVStream.codec.
The change in the lavf-mkv test is due to the native timebase (1/1000)
being used instead of the default one (1/90000), so the packets are now
sent to the crc muxer in the same order in which they are demuxed
(previously some of them got reordered because of inexact timestamp
conversion).
Anton Khirnov [Mon, 9 Jun 2014 12:21:56 +0000 (14:21 +0200)]
Remove avserver.
It has not been properly maintained for years and there is little hope
of that changing in the future.
It appears simpler to write a new replacement from scratch than
unbreaking it.
Janne Grunau [Wed, 11 Jun 2014 17:40:28 +0000 (19:40 +0200)]
mpegvideo: synchronize AVFrame pointers in ERContext fully
Since error resilience uses AVFrame pointers instead of references it
has to copy NULL pointers too. After a codec flush the last/next frame
pointers in MpegEncContext are NULL and the old pointers remaining in
ERContext are invalid. Fixes a crash in vlc for android thumbnailer.
Reported and debugged by Adrien Maglo <magsoft@videolan.org>.
Martin Storsjö [Mon, 20 Feb 2012 09:24:35 +0000 (11:24 +0200)]
x86: h264: Don't keep data in the redzone across function calls on 64 bit unix
We know that the called function (ff_chroma_inter_body_mmxext)
doesn't touch the redzone, and thus will be kept intact - thus,
this doesn't fix any bug per se.
However, valgrind's memcheck tool intentionally assumes that the
redzone is clobbered on every function call and function return
(see a long comment in valgrind/memcheck/mc_main.c). This avoids
false positives in that tool, at the cost of an extra stack pointer
adjustment.
The other alternative would be a valgrind suppression for this issue,
but that's an extra burden for everybody that wants to run libavcodec
within valgrind.
Martin Storsjö [Thu, 5 Jun 2014 11:49:14 +0000 (14:49 +0300)]
adpcm: Write the proper predictor in trellis mode in IMA QT
The actual predictor value, set by the trellis code, never
was written back into the variable that was written into
the block header. This was accidentally removed in b304244b.
This significantly improves the audio quality of the trellis
case, which was plain broken since b304244b.
Encoding IMA QT with trellis still actually gives a slightly
worse quality than without trellis, since the trellis encoder
doesn't use the exact same way of rounding as in
adpcm_ima_qt_compress_sample and adpcm_ima_qt_expand_nibble.
CC: libav-stable@libav.org Signed-off-by: Martin Storsjö <martin@martin.st>
Janne Grunau [Thu, 29 May 2014 15:04:57 +0000 (17:04 +0200)]
configure: use .altmacro for gnu as check on arm
Clang's integrated assembler (after 3.4) does not yet support the
'.altmacro' directive which is only used in arm asm. Support is planned:
http://llvm.org/bugs/show_bug.cgi?id=18918
Martin Storsjö [Thu, 5 Jun 2014 08:48:53 +0000 (11:48 +0300)]
adpcm: Avoid reading out of bounds in the IMA QT trellis encoder
This was broken in 095be4fb - samples+ch (for the previous
non-planar case) equals &samples_p[ch][0]. The confusion
probably stemmed from the IMA WAV case where it originally
was &samples[avctx->channels + ch], which was correctly
changed into &samples_p[ch][1].
CC: libav-stable@libav.org Signed-off-by: Martin Storsjö <martin@martin.st>
Martin Storsjö [Fri, 6 Jun 2014 10:59:14 +0000 (13:59 +0300)]
oggenc: Set the right AVOption size for the pref_duration option
On big endian machines, the default value set via the faulty
AVOption ended up as 2^32 times too big.
This fixes the fate-lavf-ogg test which currently is broken on
big endian machines, broken since 3831362. Since that commit,
a final zero-sized packet is written to the ogg muxer in that test,
which caused different flushing behaviour on little and big endian
depending on whether the pref_duration option was handled as it
should or not.
CC: libav-stable@libav.org Signed-off-by: Martin Storsjö <martin@martin.st>