in mpegvideo.h to avoid including mpegvideo_common.h in mpeg12.c,
because it contains motion code.
Originally committed as revision 18968 to svn://svn.ffmpeg.org/ffmpeg/trunk
As a side-effect this also gives it the correct value on e.g. PPC970FX-based
PPC64 systems, thus fixing "make test" (mp2/mp3 decoding).
Originally committed as revision 18953 to svn://svn.ffmpeg.org/ffmpeg/trunk
into the context to avoid issues when stack variables can not be aligned
reliably.
Originally committed as revision 18940 to svn://svn.ffmpeg.org/ffmpeg/trunk
idct_put for compilers/architectures that can not align stack variables that much.
This is also consistent with similar code in eatgq.c
Originally committed as revision 18927 to svn://svn.ffmpeg.org/ffmpeg/trunk
Might avoid crashes in unlikely cases, but mostly avoids ugly artefacts
for partial frames.
Originally committed as revision 18925 to svn://svn.ffmpeg.org/ffmpeg/trunk
for idct_put on architectures/compilers that have problems aligning on the stack.
Originally committed as revision 18924 to svn://svn.ffmpeg.org/ffmpeg/trunk
This ensures that the MMX loop filter is always bitexact with the C version.
Patch by Haruhiko Yamagata <h.yamagata _a_ nifty com>
Originally committed as revision 18923 to svn://svn.ffmpeg.org/ffmpeg/trunk
changeing calls to sub_left_prediction such that
the source frame data pointers are no longer offset to skip the initial
elements. This keeps them properly aligned for the subsequent call to diff_bytes.
Changes encoding offset to compensate.
Mailing list thread:
https://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2009-May/069952.html
Originally committed as revision 18919 to svn://svn.ffmpeg.org/ffmpeg/trunk
in the source data array to start encoding.
Change all calls to pass zero for this new argument for equivalent behavior.
This is a restructure to support a forthcoming bug fix.
Originally committed as revision 18918 to svn://svn.ffmpeg.org/ffmpeg/trunk
MLP DSP functions x86-optimized.
12.59% overall speedup in x86_32
9.98% overall speedup in x86_64
compared to gcc 4.3.3
It breaks on x86_64 with PIC because the MANGLE macro appends (%%rip), which
is not wanted for labels.
Originally committed as revision 18904 to svn://svn.ffmpeg.org/ffmpeg/trunk
12.59% overall speedup in x86_32
9.98% overall speedup in x86_64
compared to gcc 4.3.3
Originally committed as revision 18903 to svn://svn.ffmpeg.org/ffmpeg/trunk
it is aligned as necessary for DSPContext's idct_put.
Just aligning it on the stack would have been possible but less reliable
and without any real benefit.
Originally committed as revision 18896 to svn://svn.ffmpeg.org/ffmpeg/trunk
Patch by Martin Storsjö: martin (remove that) at martin (oh and this too) st
Originally committed as revision 18880 to svn://svn.ffmpeg.org/ffmpeg/trunk
Now the sign of audio samples in ffvorbis output is the same as in original
uncompressed audio file and this also allows the use of tiny_psnr to compare
ffvorbis with libvorbis/tremor.
Originally committed as revision 18878 to svn://svn.ffmpeg.org/ffmpeg/trunk
