platform/upstream/libav.git
14 years agoVP8: partially inline decode_block_coeffs
Jason Garrett-Glaser [Wed, 4 Aug 2010 02:23:25 +0000 (02:23 +0000)]
VP8: partially inline decode_block_coeffs
Avoids a function call in the case of empty DCT blocks (most of the time).

Originally committed as revision 24691 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix 100L in r24689
Jason Garrett-Glaser [Wed, 4 Aug 2010 01:40:58 +0000 (01:40 +0000)]
Fix 100L in r24689
Accidentally committed some timing code.

Originally committed as revision 24690 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: simplify decode_block_coeffs to avoid having to track nonzero coeffs
Jason Garrett-Glaser [Wed, 4 Aug 2010 01:38:08 +0000 (01:38 +0000)]
VP8: simplify decode_block_coeffs to avoid having to track nonzero coeffs
Slightly faster.

Originally committed as revision 24689 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetics: add missing space to options.c
Jason Garrett-Glaser [Wed, 4 Aug 2010 00:08:16 +0000 (00:08 +0000)]
Cosmetics: add missing space to options.c

Originally committed as revision 24688 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: slightly faster DCT coefficient probability update
Jason Garrett-Glaser [Tue, 3 Aug 2010 23:21:47 +0000 (23:21 +0000)]
VP8: slightly faster DCT coefficient probability update

Originally committed as revision 24687 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoARM: update struct offsets
Måns Rullgård [Tue, 3 Aug 2010 22:29:38 +0000 (22:29 +0000)]
ARM: update struct offsets

Originally committed as revision 24686 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove cavs dsp functions to their own struct
Måns Rullgård [Tue, 3 Aug 2010 20:59:00 +0000 (20:59 +0000)]
Move cavs dsp functions to their own struct

Originally committed as revision 24685 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSilence unused function warnings in vp56.h
Jason Garrett-Glaser [Tue, 3 Aug 2010 11:35:41 +0000 (11:35 +0000)]
Silence unused function warnings in vp56.h

Originally committed as revision 24684 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: make another RAC call branchy
Jason Garrett-Glaser [Tue, 3 Aug 2010 11:34:24 +0000 (11:34 +0000)]
VP8: make another RAC call branchy
1-2 clocks faster.

Originally committed as revision 24683 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP5/6/8: add one inline missed in r24677
Jason Garrett-Glaser [Tue, 3 Aug 2010 11:21:22 +0000 (11:21 +0000)]
VP5/6/8: add one inline missed in r24677

Originally committed as revision 24682 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: unroll partition type decoding tree
Jason Garrett-Glaser [Tue, 3 Aug 2010 11:10:58 +0000 (11:10 +0000)]
VP8: unroll partition type decoding tree
~34% faster partition type decoding.

Originally committed as revision 24681 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: unroll splitmv decoding tree
Jason Garrett-Glaser [Tue, 3 Aug 2010 10:37:14 +0000 (10:37 +0000)]
VP8: unroll splitmv decoding tree
Much faster splitmv mode decoding.

Originally committed as revision 24680 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: unroll MB mode decoding tree
Jason Garrett-Glaser [Tue, 3 Aug 2010 10:24:28 +0000 (10:24 +0000)]
VP8: unroll MB mode decoding tree
~50% faster MB mode decoding, plus eliminate a costly switch.

Originally committed as revision 24679 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMake avfilter_copy_picref_props() copy w and h from src to dst.
Stefano Sabatini [Tue, 3 Aug 2010 09:02:16 +0000 (09:02 +0000)]
Make avfilter_copy_picref_props() copy w and h from src to dst.

Originally committed as revision 24678 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP5/6/8: tweak some arithcoder inlining
Jason Garrett-Glaser [Tue, 3 Aug 2010 08:06:08 +0000 (08:06 +0000)]
VP5/6/8: tweak some arithcoder inlining
Always inline the arithmetic coder, except in the case of header-parsing stuff,
in which case don't inline it at all to save code size.

Originally committed as revision 24677 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agovorbisdec: Return AVERROR(ENOMEM) on malloc() failure.
Alex Converse [Tue, 3 Aug 2010 00:27:54 +0000 (00:27 +0000)]
vorbisdec: Return AVERROR(ENOMEM) on malloc() failure.

This is especially important because classifs can be very large.

Originally committed as revision 24676 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agovorbisdec: Prevent a potential integer overflow.
Alex Converse [Tue, 3 Aug 2010 00:25:06 +0000 (00:25 +0000)]
vorbisdec: Prevent a potential integer overflow.

If sizeof uint_fast8_t > 1 and sizeof size_t <= 4, the expression that mallocs
classifs  is susceptible to integer overflow.

Originally committed as revision 24675 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoThe VP8 decoder does not depend on cabac.o.
Diego Biurrun [Mon, 2 Aug 2010 23:49:49 +0000 (23:49 +0000)]
The VP8 decoder does not depend on cabac.o.

Originally committed as revision 24674 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoThe VP8 decoder does not depend on vp56.o and vp56data.o.
Diego Biurrun [Mon, 2 Aug 2010 23:23:27 +0000 (23:23 +0000)]
The VP8 decoder does not depend on vp56.o and vp56data.o.

Originally committed as revision 24673 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove redundant h264pred.o from VP8 objects list.
Diego Biurrun [Mon, 2 Aug 2010 23:16:52 +0000 (23:16 +0000)]
Remove redundant h264pred.o from VP8 objects list.
It is already selected through the H264DSP dependency of VP8.

Originally committed as revision 24672 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: eliminate a dereference in coefficient decoding
Jason Garrett-Glaser [Mon, 2 Aug 2010 22:48:38 +0000 (22:48 +0000)]
VP8: eliminate a dereference in coefficient decoding

Originally committed as revision 24671 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: much faster DC transform handling
Jason Garrett-Glaser [Mon, 2 Aug 2010 20:57:03 +0000 (20:57 +0000)]
VP8: much faster DC transform handling
A lot of the time the DC block is empty: don't do the WHT in this case.
A lot of the rest of the time, there's only one coefficient: make a special
DC-only transform for that case.
When the block is empty, don't incorrectly mark luma DCT blocks as having DC
coefficients.

Originally committed as revision 24670 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agovorbisdec: change a uint_fast32_t to 'unsigned', fix llvm-gcc build
Måns Rullgård [Mon, 2 Aug 2010 20:35:50 +0000 (20:35 +0000)]
vorbisdec: change a uint_fast32_t to 'unsigned', fix llvm-gcc build

Originally committed as revision 24669 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: move zeroing of luma DC block into the WHT
Jason Garrett-Glaser [Mon, 2 Aug 2010 20:18:09 +0000 (20:18 +0000)]
VP8: move zeroing of luma DC block into the WHT
Lets us do the zeroing in asm instead of C.
Also makes it consistent with the way the regular iDCT code does it.

Originally committed as revision 24668 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agolavfi-regression: trim \r line ends from showfiltfmts output
Måns Rullgård [Mon, 2 Aug 2010 19:44:58 +0000 (19:44 +0000)]
lavfi-regression: trim \r line ends from showfiltfmts output

This is required when cross-testing on dos-like targets.

Originally committed as revision 24667 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd -Wno-parentheses to CFLAGS.
Diego Biurrun [Mon, 2 Aug 2010 17:09:55 +0000 (17:09 +0000)]
Add -Wno-parentheses to CFLAGS.
Without this flag -Wall generates warnings that nobody wants to see fixed or
worked around. Disabling those warnings is thus the best solution.

Originally committed as revision 24666 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSort lavfi pixdesc tests.
Stefano Sabatini [Mon, 2 Aug 2010 11:56:51 +0000 (11:56 +0000)]
Sort lavfi pixdesc tests.

Increase readability and robustness, as the test result is not going
to differ if the order of the pixfmts codes changes.

Originally committed as revision 24665 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoonly store intra prediction modes on the boundary for keyframes, not as a plane.
Pascal Massimino [Mon, 2 Aug 2010 09:44:53 +0000 (09:44 +0000)]
only store intra prediction modes on the boundary for keyframes, not as a plane.
inter-frame behaviour unchanged.

Originally committed as revision 24664 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix doxygen comment.
Benoit Fouet [Mon, 2 Aug 2010 07:28:35 +0000 (07:28 +0000)]
Fix doxygen comment.

Originally committed as revision 24663 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoChange a doxy comment to a normal one.
Benoit Fouet [Mon, 2 Aug 2010 07:26:27 +0000 (07:26 +0000)]
Change a doxy comment to a normal one.

Originally committed as revision 24662 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix doxygen comments.
Benoit Fouet [Mon, 2 Aug 2010 07:25:25 +0000 (07:25 +0000)]
Fix doxygen comments.
/**\u2264 => /**<

Originally committed as revision 24661 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoIssue a warning when fed with misformatted one-line doxygen comments.
Benoit Fouet [Mon, 2 Aug 2010 07:23:47 +0000 (07:23 +0000)]
Issue a warning when fed with misformatted one-line doxygen comments.

Originally committed as revision 24660 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoVP8: simplify token_prob handling
Jason Garrett-Glaser [Mon, 2 Aug 2010 05:20:38 +0000 (05:20 +0000)]
VP8: simplify token_prob handling
~1.5% faster decode_block_coeffs

Originally committed as revision 24659 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agolavfi-regression: use different temp file names for each pixfmt test
Måns Rullgård [Sun, 1 Aug 2010 23:34:48 +0000 (23:34 +0000)]
lavfi-regression: use different temp file names for each pixfmt test

Originally committed as revision 24658 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd lavfi-pixfmts LE tests.
Stefano Sabatini [Sun, 1 Aug 2010 23:20:46 +0000 (23:20 +0000)]
Add lavfi-pixfmts LE tests.

The corresponding lavfi-pixfmts BE tests are not yet added, as there
are some bugs in the scaler (scaling rgba, argb, bgra, abgr, yuva420p)
which result in differences with the LE reference, and I cannot
visually check the generated files on BE.

Originally committed as revision 24657 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoprevent access to vp8_coeff_band[16]
Pascal Massimino [Sun, 1 Aug 2010 23:20:06 +0000 (23:20 +0000)]
prevent access to vp8_coeff_band[16]

Originally committed as revision 24656 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoReindent.
Stefano Sabatini [Sun, 1 Aug 2010 23:04:16 +0000 (23:04 +0000)]
Reindent.

Originally committed as revision 24655 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSplit lavfi pixfmts test.
Stefano Sabatini [Sun, 1 Aug 2010 23:04:13 +0000 (23:04 +0000)]
Split lavfi pixfmts test.

Introduce the function do_lavfi_pixfmts(), and use it for generating a
pixfmts test for each different filter.

Originally committed as revision 24654 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoImplement set_ne_test_deps() and use if for the lavfi pixdesc test.
Stefano Sabatini [Sun, 1 Aug 2010 23:04:10 +0000 (23:04 +0000)]
Implement set_ne_test_deps() and use if for the lavfi pixdesc test.

Originally committed as revision 24653 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix suncc ident string (hopefully)
Måns Rullgård [Sun, 1 Aug 2010 16:13:12 +0000 (16:13 +0000)]
Fix suncc ident string (hopefully)

Originally committed as revision 24652 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agofate: fix non-standard use of bc
Måns Rullgård [Sun, 1 Aug 2010 15:28:42 +0000 (15:28 +0000)]
fate: fix non-standard use of bc

Originally committed as revision 24651 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agofate: fix signal name translation
Måns Rullgård [Sun, 1 Aug 2010 12:29:03 +0000 (12:29 +0000)]
fate: fix signal name translation

Originally committed as revision 24650 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agofate: make tar command configurable
Måns Rullgård [Sun, 1 Aug 2010 12:29:00 +0000 (12:29 +0000)]
fate: make tar command configurable

The 'tar' variable should be set to a command writing a tar archive
of the named files to stdout, typically "tar c" or "tar cf -"

Originally committed as revision 24649 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSimplify lavfi-pixfmts test script
Måns Rullgård [Sun, 1 Aug 2010 11:34:59 +0000 (11:34 +0000)]
Simplify lavfi-pixfmts test script

Originally committed as revision 24648 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agolavfi-showfiltfmts: print one format per line
Måns Rullgård [Sun, 1 Aug 2010 11:34:56 +0000 (11:34 +0000)]
lavfi-showfiltfmts: print one format per line

Originally committed as revision 24647 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove temporary files in lavfi-pixfmts test
Måns Rullgård [Sun, 1 Aug 2010 10:47:41 +0000 (10:47 +0000)]
Remove temporary files in lavfi-pixfmts test

Originally committed as revision 24646 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix fate-lavfi-pixfmts test cross-compilation.
Stefano Sabatini [Sun, 1 Aug 2010 10:27:04 +0000 (10:27 +0000)]
Fix fate-lavfi-pixfmts test cross-compilation.

Add the lavfi-showfiltfmts dependency in the Makefile, and correctly
use the $target_exec and $target_path variables for invoking the
lavfi-showfiltfmts tool.

Originally committed as revision 24645 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUse the ffmpeg specified in $ffmpeg in the pixfmts lavfi test,
Stefano Sabatini [Sun, 1 Aug 2010 10:27:00 +0000 (10:27 +0000)]
Use the ffmpeg specified in $ffmpeg in the pixfmts lavfi test,
otherwise the test will be running whatever ffmpeg is installed on the
host system.

Originally committed as revision 24644 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoPut the filter name before the pixel format name in the lavfi pixfmts
Stefano Sabatini [Sun, 1 Aug 2010 00:46:11 +0000 (00:46 +0000)]
Put the filter name before the pixel format name in the lavfi pixfmts
test output files, and add a prefix with the name of the test.
Make per-filter grouping of the generated output files easier, which
is more useful than per-pixel-format grouping.

Originally committed as revision 24643 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoIntroduce and use a variable $output in the lavfi pixfmts test code.
Stefano Sabatini [Sun, 1 Aug 2010 00:46:08 +0000 (00:46 +0000)]
Introduce and use a variable $output in the lavfi pixfmts test code.
Consistent with the lavfi pixdesc test code, and slightly improve
readability.

Originally committed as revision 24642 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSplit the lavfi pixfmts tests in _le and _be, this is required as the
Stefano Sabatini [Sun, 1 Aug 2010 00:46:06 +0000 (00:46 +0000)]
Split the lavfi pixfmts tests in _le and _be, this is required as the
test results and references depend on machine endianess.

Originally committed as revision 24641 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRename the not yet enabled test lavfi_pix_fmts to pixfmts, which is
Stefano Sabatini [Sun, 1 Aug 2010 00:46:03 +0000 (00:46 +0000)]
Rename the not yet enabled test lavfi_pix_fmts to pixfmts, which is
simpler and consistent with the names of the other lavfi tests.

Originally committed as revision 24640 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd WebM to the Matroska demuxer name.
Alex Converse [Sun, 1 Aug 2010 00:30:37 +0000 (00:30 +0000)]
Add WebM to the Matroska demuxer name.

Originally committed as revision 24639 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUse word-writing instead of dword-writing (with two cached but otherwise
Ronald S. Bultje [Sat, 31 Jul 2010 23:13:15 +0000 (23:13 +0000)]
Use word-writing instead of dword-writing (with two cached but otherwise
unchanged bytes) in the horizontal simple loopfilter. This makes the filter
quite a bit faster in itself (~30 cycles less on Core1), probably mostly
because we don't need a complex 4x4 transpose, but only a simple byte
interleave. Also allows using pextrw on SSE4, which speeds up even more
(e.g. 25% faster on Core i7).

Originally committed as revision 24638 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agodca: fix dynrange coefficient in xch
Nick Brereton [Sat, 31 Jul 2010 22:15:59 +0000 (22:15 +0000)]
dca: fix dynrange coefficient in xch

Patch by Nick Brereton

Originally committed as revision 24637 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoswscale-test: merge declaration and initialization
Ramiro Polla [Sat, 31 Jul 2010 22:02:55 +0000 (22:02 +0000)]
swscale-test: merge declaration and initialization

Originally committed as revision 31879 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agocosmetics: rename output_* to write_*
Justin Ruggles [Sat, 31 Jul 2010 21:14:03 +0000 (21:14 +0000)]
cosmetics: rename output_* to write_*

Originally committed as revision 24634 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSimplify verbatim mode fallback by checking the frame size before writing.
Justin Ruggles [Sat, 31 Jul 2010 21:08:53 +0000 (21:08 +0000)]
Simplify verbatim mode fallback by checking the frame size before writing.

Originally committed as revision 24632 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoChange max_framesize for small final frame.
Justin Ruggles [Sat, 31 Jul 2010 20:52:33 +0000 (20:52 +0000)]
Change max_framesize for small final frame.

Originally committed as revision 24631 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCalculate an exact frame size before writing. Now the buffer size requirements
Justin Ruggles [Sat, 31 Jul 2010 20:32:12 +0000 (20:32 +0000)]
Calculate an exact frame size before writing.  Now the buffer size requirements
can be known exactly, so larger frame sizes can be safely encoded without buffer
overwrite.

Originally committed as revision 24630 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: rename find_subblock_rice_params() to find_subframe_rice_params()
Justin Ruggles [Sat, 31 Jul 2010 20:19:07 +0000 (20:19 +0000)]
cosmetics: rename find_subblock_rice_params() to find_subframe_rice_params()

Originally committed as revision 24629 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoReduce number of input parameters to find_subblock_rice_params().
Justin Ruggles [Sat, 31 Jul 2010 20:17:59 +0000 (20:17 +0000)]
Reduce number of input parameters to find_subblock_rice_params().

Originally committed as revision 24628 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years ago10l: fix bit count for frame header
Justin Ruggles [Sat, 31 Jul 2010 18:48:48 +0000 (18:48 +0000)]
10l: fix bit count for frame header

Originally committed as revision 24627 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSimplify fallback to verbatim mode encoding.
Justin Ruggles [Sat, 31 Jul 2010 17:23:29 +0000 (17:23 +0000)]
Simplify fallback to verbatim mode encoding.

Originally committed as revision 24626 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: line wrap and vertical alignment
Justin Ruggles [Sat, 31 Jul 2010 17:07:24 +0000 (17:07 +0000)]
cosmetics: line wrap and vertical alignment

Originally committed as revision 24625 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine calc_rice_params_fixed() and calc_rice_params_lpc() into a single
Justin Ruggles [Sat, 31 Jul 2010 17:06:10 +0000 (17:06 +0000)]
Combine calc_rice_params_fixed() and calc_rice_params_lpc() into a single
function.

Originally committed as revision 24624 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agofate: translate exit status to signal name
Måns Rullgård [Sat, 31 Jul 2010 16:53:09 +0000 (16:53 +0000)]
fate: translate exit status to signal name

Originally committed as revision 24623 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove unneeded wrapper function.
Justin Ruggles [Sat, 31 Jul 2010 16:50:27 +0000 (16:50 +0000)]
Remove unneeded wrapper function.

Originally committed as revision 24622 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove duplicate code by adding a flag for encoding in verbatim mode.
Justin Ruggles [Sat, 31 Jul 2010 16:46:32 +0000 (16:46 +0000)]
Remove duplicate code by adding a flag for encoding in verbatim mode.

Originally committed as revision 24621 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUse av_fill_image_pointers/linesizes in place of ff_fill_pointer/linesize,
Stefano Sabatini [Sat, 31 Jul 2010 16:46:20 +0000 (16:46 +0000)]
Use av_fill_image_pointers/linesizes in place of ff_fill_pointer/linesize,
and drop the the ff_ variants at the next major bump.

Originally committed as revision 24620 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoEstimate frame size during encoding.
Justin Ruggles [Sat, 31 Jul 2010 16:37:23 +0000 (16:37 +0000)]
Estimate frame size during encoding.

Originally committed as revision 24619 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove x86/mmx.h. It is not used anymore and has been deprecated for years.
Vitor Sessak [Sat, 31 Jul 2010 16:20:45 +0000 (16:20 +0000)]
Remove x86/mmx.h. It is not used anymore and has been deprecated for years.

Originally committed as revision 24618 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix compilation with --disable-yasm. 10l to me.
Vitor Sessak [Sat, 31 Jul 2010 16:17:54 +0000 (16:17 +0000)]
Fix compilation with --disable-yasm. 10l to me.

Originally committed as revision 24617 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd protocols.texi.
Stefano Sabatini [Sat, 31 Jul 2010 15:45:29 +0000 (15:45 +0000)]
Add protocols.texi.

Originally committed as revision 24616 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoConvert deinterlacing MMX code to YASM
Vitor Sessak [Sat, 31 Jul 2010 14:50:51 +0000 (14:50 +0000)]
Convert deinterlacing MMX code to YASM

Originally committed as revision 24615 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoconfigure: set subarch for ARM
Måns Rullgård [Sat, 31 Jul 2010 08:51:39 +0000 (08:51 +0000)]
configure: set subarch for ARM

Originally committed as revision 24614 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years ago100l: av_freep() needs the address of the pointer
Aurelien Jacobs [Fri, 30 Jul 2010 23:41:12 +0000 (23:41 +0000)]
100l: av_freep() needs the address of the pointer

Originally committed as revision 24613 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: indentation
Justin Ruggles [Fri, 30 Jul 2010 21:33:51 +0000 (21:33 +0000)]
cosmetics: indentation

Originally committed as revision 24612 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine output_subframe() and output_subframes().
Justin Ruggles [Fri, 30 Jul 2010 21:32:19 +0000 (21:32 +0000)]
Combine output_subframe() and output_subframes().

Originally committed as revision 24611 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove unneeded variable.
Justin Ruggles [Fri, 30 Jul 2010 21:27:00 +0000 (21:27 +0000)]
Remove unneeded variable.

Originally committed as revision 24610 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: indentation
Justin Ruggles [Fri, 30 Jul 2010 21:07:21 +0000 (21:07 +0000)]
cosmetics: indentation

Originally committed as revision 24609 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine output_subframe_verbatim() and output_subframe_lpc().
Justin Ruggles [Fri, 30 Jul 2010 21:06:38 +0000 (21:06 +0000)]
Combine output_subframe_verbatim() and output_subframe_lpc().

Originally committed as revision 24608 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine and simplify output_residual() and output_subframe_lpc().
Justin Ruggles [Fri, 30 Jul 2010 20:53:02 +0000 (20:53 +0000)]
Combine and simplify output_residual() and output_subframe_lpc().

Originally committed as revision 24607 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: reindent
Justin Ruggles [Fri, 30 Jul 2010 20:29:03 +0000 (20:29 +0000)]
cosmetics: reindent

Originally committed as revision 24606 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine and simplify output_subframe_fixed() and output_subframe_lpc().
Justin Ruggles [Fri, 30 Jul 2010 20:28:46 +0000 (20:28 +0000)]
Combine and simplify output_subframe_fixed() and output_subframe_lpc().

Originally committed as revision 24605 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCombine and simplify output_subframe_constant() and output_subframe_verbatim().
Justin Ruggles [Fri, 30 Jul 2010 20:24:15 +0000 (20:24 +0000)]
Combine and simplify output_subframe_constant() and output_subframe_verbatim().

Originally committed as revision 24604 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoPass FlacSubframe to output_subframe_* instead of channel number.
Justin Ruggles [Fri, 30 Jul 2010 20:09:14 +0000 (20:09 +0000)]
Pass FlacSubframe to output_subframe_* instead of channel number.

Originally committed as revision 24603 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: change FlacEncodeContext variable name from ctx to s in several
Justin Ruggles [Fri, 30 Jul 2010 19:40:18 +0000 (19:40 +0000)]
cosmetics: change FlacEncodeContext variable name from ctx to s in several
places for consistency.

Originally committed as revision 24602 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSet coded_frame->pts in the FLAC encoder
Justin Ruggles [Fri, 30 Jul 2010 19:18:48 +0000 (19:18 +0000)]
Set coded_frame->pts in the FLAC encoder

Originally committed as revision 24601 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoDo not need to set coded_frame->key_frame = 1 because it is already set in
Justin Ruggles [Fri, 30 Jul 2010 19:02:59 +0000 (19:02 +0000)]
Do not need to set coded_frame->key_frame = 1 because it is already set in
avcodec_alloc_frame().

Originally committed as revision 24600 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove debug logging of compression options to a single function.
Justin Ruggles [Fri, 30 Jul 2010 18:59:27 +0000 (18:59 +0000)]
Move debug logging of compression options to a single function.

Originally committed as revision 24599 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd 2 failed memory allocation checks
Justin Ruggles [Fri, 30 Jul 2010 18:34:39 +0000 (18:34 +0000)]
Add 2 failed memory allocation checks

Originally committed as revision 24598 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: pretty-print flacenc.c
Justin Ruggles [Fri, 30 Jul 2010 18:30:09 +0000 (18:30 +0000)]
cosmetics: pretty-print flacenc.c

Originally committed as revision 24597 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove mostly unnecessary rtpdec_*.h files, store the declarations in one file
Martin Storsjö [Fri, 30 Jul 2010 12:04:27 +0000 (12:04 +0000)]
Remove mostly unnecessary rtpdec_*.h files, store the declarations in one file

Originally committed as revision 24596 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMake config_input() return AVERROR(EINVAL) in place of -1.
Stefano Sabatini [Fri, 30 Jul 2010 11:56:01 +0000 (11:56 +0000)]
Make config_input() return AVERROR(EINVAL) in place of -1.

Originally committed as revision 24595 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCompute the max pixel step for each plane, and use it in place of
Stefano Sabatini [Fri, 30 Jul 2010 11:33:31 +0000 (11:33 +0000)]
Compute the max pixel step for each plane, and use it in place of
hardcoding that value in a switch.

More compact and correct.

Originally committed as revision 24594 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd an entry for the AVFilterBuffer change of r24592.
Stefano Sabatini [Fri, 30 Jul 2010 09:06:03 +0000 (09:06 +0000)]
Add an entry for the AVFilterBuffer change of r24592.

Originally committed as revision 24593 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoResize AVFilterBuffer to hold 8 planes/linesizes worth of data.
S.N. Hemanth Meenakshisundaram [Fri, 30 Jul 2010 08:42:25 +0000 (08:42 +0000)]
Resize AVFilterBuffer to hold 8 planes/linesizes worth of data.

This is required for making AVFilterBuffer useful for storing audio
data.

Patch by S.N. Hemanth Meenakshisundaram revert <ude.dscu@skaneems>.

Originally committed as revision 24592 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRename the av_fill_image_linesize() formal parameter linesize to
Stefano Sabatini [Fri, 30 Jul 2010 08:42:22 +0000 (08:42 +0000)]
Rename the av_fill_image_linesize() formal parameter linesize to
linesizes, for consistency with the function declaration.

Originally committed as revision 24591 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoReimplement av_fill_image_pointers() using the information stored in
Stefano Sabatini [Fri, 30 Jul 2010 08:42:19 +0000 (08:42 +0000)]
Reimplement av_fill_image_pointers() using the information stored in
the pixdescs.

The new implementation is more compact, and does not need to be
updated at each pixel format addition.

Originally committed as revision 24590 to svn://svn.ffmpeg.org/ffmpeg/trunk