Statistics
| Branch: | Revision:

ffmpeg @ 476be414

# Date Author Comment
476be414 08/03/2010 11:34 AM Jason Garrett-Glaser

VP8: make another RAC call branchy
1-2 clocks faster.

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

8b9b5e08 08/03/2010 11:21 AM Jason Garrett-Glaser

VP5/6/8: add one inline missed in r24677

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

0908f1b9 08/03/2010 11:10 AM Jason Garrett-Glaser

VP8: unroll partition type decoding tree
~34% faster partition type decoding.

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

c5dec7f1 08/03/2010 10:37 AM Jason Garrett-Glaser

VP8: unroll splitmv decoding tree
Much faster splitmv mode decoding.

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

23117d69 08/03/2010 10:24 AM Jason Garrett-Glaser

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

be665c7d 08/03/2010 09:02 AM Stefano Sabatini

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

cd2769c1 08/03/2010 08:06 AM Jason Garrett-Glaser

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

55aa55f2 08/03/2010 12:27 AM Alex Converse

vorbisdec: Return AVERROR 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

366d9190 08/03/2010 12:25 AM Alex Converse

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

83abdf5f 08/02/2010 11:49 PM Diego Biurrun

The VP8 decoder does not depend on cabac.o.

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

efbc4559 08/02/2010 11:23 PM Diego Biurrun

The VP8 decoder does not depend on vp56.o and vp56data.o.

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

2e2ca506 08/02/2010 11:16 PM Diego Biurrun

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

370b622a 08/02/2010 10:48 PM Jason Garrett-Glaser

VP8: eliminate a dereference in coefficient decoding

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

f311208c 08/02/2010 08:57 PM Jason Garrett-Glaser

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...

c934562c 08/02/2010 08:35 PM Måns Rullgård

vorbisdec: change a uint_fast32_t to 'unsigned', fix llvm-gcc build

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

827d43bb 08/02/2010 08:18 PM Jason Garrett-Glaser

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

42907c6a 08/02/2010 07:44 PM Måns Rullgård

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

440e3b2f 08/02/2010 05:09 PM Diego Biurrun

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

546017fc 08/02/2010 11:56 AM Stefano Sabatini

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

d2840fa4 08/02/2010 09:44 AM Pascal Massimino

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

eea72ee8 08/02/2010 07:28 AM Benoit Fouet

Fix doxygen comment.

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

497d7991 08/02/2010 07:26 AM Benoit Fouet

Change a doxy comment to a normal one.

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

1aa15491 08/02/2010 07:25 AM Benoit Fouet

Fix doxygen comments.
/**\u2264 => /**<

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

3ea4d942 08/02/2010 07:23 AM Benoit Fouet

Issue a warning when fed with misformatted one-line doxygen comments.

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

10bf2eeb 08/02/2010 05:20 AM Jason Garrett-Glaser

VP8: simplify token_prob handling
~1.5% faster decode_block_coeffs

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

9c87c037 08/01/2010 11:34 PM Måns Rullgård

lavfi-regression: use different temp file names for each pixfmt test

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

213a4cbb 08/01/2010 11:20 PM Stefano Sabatini

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....

c22b4468 08/01/2010 11:20 PM Pascal Massimino

prevent access to vp8_coeff_band16

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

04307697 08/01/2010 11:04 PM Stefano Sabatini

Reindent.

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

73177b6c 08/01/2010 11:04 PM Stefano Sabatini

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

9c261f83 08/01/2010 11:04 PM Stefano Sabatini

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

448f5b83 08/01/2010 04:13 PM Måns Rullgård

Fix suncc ident string (hopefully)

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

61a187b0 08/01/2010 03:28 PM Måns Rullgård

fate: fix non-standard use of bc

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

5b60c293 08/01/2010 12:29 PM Måns Rullgård

fate: fix signal name translation

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

8b0816cb 08/01/2010 12:29 PM Måns Rullgård

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

7e7fc4e9 08/01/2010 11:34 AM Måns Rullgård

Simplify lavfi-pixfmts test script

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

27971664 08/01/2010 11:34 AM Måns Rullgård

lavfi-showfiltfmts: print one format per line

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

79ec38d1 08/01/2010 10:47 AM Måns Rullgård

Remove temporary files in lavfi-pixfmts test

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

43d2c6a5 08/01/2010 10:27 AM Stefano Sabatini

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

2b346e04 08/01/2010 10:27 AM Stefano Sabatini

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

0a87d541 08/01/2010 12:46 AM Stefano Sabatini

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

65e17eae 08/01/2010 12:46 AM Stefano Sabatini

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

fc25373b 08/01/2010 12:46 AM Stefano Sabatini

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

42bc7c8a 08/01/2010 12:46 AM Stefano Sabatini

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

f7cf0f31 08/01/2010 12:30 AM Alex Converse

Add WebM to the Matroska demuxer name.

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

6341838f 07/31/2010 11:13 PM Ronald S. Bultje

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...

ace7f813 07/31/2010 10:15 PM Nick Brereton

dca: fix dynrange coefficient in xch

Patch by Nick Brereton

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

e07b1939 07/31/2010 10:02 PM Ramiro Polla

swscale-test: merge declaration and initialization

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

6768beb4 07/31/2010 09:14 PM Justin Ruggles

cosmetics: rename output_* to write_*

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

3d2cd42f 07/31/2010 09:08 PM Justin Ruggles

Simplify verbatim mode fallback by checking the frame size before writing.

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

2249a7f3 07/31/2010 08:52 PM Justin Ruggles

Change max_framesize for small final frame.

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

7c29a5de 07/31/2010 08:32 PM Justin Ruggles

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

675eb677 07/31/2010 08:19 PM Justin Ruggles

cosmetics: rename find_subblock_rice_params() to find_subframe_rice_params()

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

089c18f3 07/31/2010 08:17 PM Justin Ruggles

Reduce number of input parameters to find_subblock_rice_params().

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

d309f019 07/31/2010 06:48 PM Justin Ruggles

10l: fix bit count for frame header

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

b22f9d65 07/31/2010 05:23 PM Justin Ruggles

Simplify fallback to verbatim mode encoding.

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

45e29087 07/31/2010 05:07 PM Justin Ruggles

cosmetics: line wrap and vertical alignment

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

8fbb8d31 07/31/2010 05:06 PM Justin Ruggles

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

4e705a68 07/31/2010 04:53 PM Måns Rullgård

fate: translate exit status to signal name

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

43b3273e 07/31/2010 04:50 PM Justin Ruggles

Remove unneeded wrapper function.

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

0f5cc12c 07/31/2010 04:46 PM Justin Ruggles

Remove duplicate code by adding a flag for encoding in verbatim mode.

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

05236ed7 07/31/2010 04:46 PM Stefano Sabatini

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

93d65e3d 07/31/2010 04:37 PM Justin Ruggles

Estimate frame size during encoding.

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

fa738b3a 07/31/2010 04:20 PM Vitor Sessak

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

0913a92a 07/31/2010 04:17 PM Vitor Sessak

Fix compilation with --disable-yasm. 10l to me.

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

1de4cfe6 07/31/2010 03:45 PM Stefano Sabatini

Add protocols.texi.

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

de4bc44a 07/31/2010 02:50 PM Vitor Sessak

Convert deinterlacing MMX code to YASM

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

c2eae137 07/31/2010 08:51 AM Måns Rullgård

configure: set subarch for ARM

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

f138fa78 07/30/2010 11:41 PM Aurelien Jacobs

100l: av_freep() needs the address of the pointer

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

eb96f169 07/30/2010 09:33 PM Justin Ruggles

cosmetics: indentation

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

520effb9 07/30/2010 09:32 PM Justin Ruggles

Combine output_subframe() and output_subframes().

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

fb443a6f 07/30/2010 09:27 PM Justin Ruggles

Remove unneeded variable.

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

724f3471 07/30/2010 09:07 PM Justin Ruggles

cosmetics: indentation

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

6d263f41 07/30/2010 09:06 PM Justin Ruggles

Combine output_subframe_verbatim() and output_subframe_lpc().

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

6266cfb4 07/30/2010 08:53 PM Justin Ruggles

Combine and simplify output_residual() and output_subframe_lpc().

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

fc001d6c 07/30/2010 08:29 PM Justin Ruggles

cosmetics: reindent

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

175a5711 07/30/2010 08:28 PM Justin Ruggles

Combine and simplify output_subframe_fixed() and output_subframe_lpc().

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

89912b5f 07/30/2010 08:24 PM Justin Ruggles

Combine and simplify output_subframe_constant() and output_subframe_verbatim().

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

1f1c05b0 07/30/2010 08:09 PM Justin Ruggles

Pass FlacSubframe to output_subframe_* instead of channel number.

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

15f537ed 07/30/2010 07:40 PM Justin Ruggles

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

871a24f3 07/30/2010 07:18 PM Justin Ruggles

Set coded_frame->pts in the FLAC encoder

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

27e34188 07/30/2010 07:02 PM Justin Ruggles

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

7fe0c7c8 07/30/2010 06:59 PM Justin Ruggles

Move debug logging of compression options to a single function.

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

e08ec714 07/30/2010 06:34 PM Justin Ruggles

Add 2 failed memory allocation checks

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

e35b689e 07/30/2010 06:30 PM Justin Ruggles

cosmetics: pretty-print flacenc.c

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

965a3ddb 07/30/2010 12:04 PM Martin Storsjö

Remove mostly unnecessary rtpdec_*.h files, store the declarations in one file

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

d6bb31a4 07/30/2010 11:56 AM Stefano Sabatini

Make config_input() return AVERROR in place of -1.

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

b59d12b1 07/30/2010 11:33 AM Stefano Sabatini

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

92dd4959 07/30/2010 09:06 AM Stefano Sabatini

Add an entry for the AVFilterBuffer change of r24592.

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

56b5e9d5 07/30/2010 08:42 AM S.N. Hemanth Meenakshisunda...

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

ecc7bfe3 07/30/2010 08:42 AM Stefano Sabatini

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

7d8d18de 07/30/2010 08:42 AM Stefano Sabatini

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

46b1d17a 07/30/2010 08:42 AM Stefano Sabatini

Make av_fill_image_linesizes() return a meaningful error core rather
than -1.

Originally committed as revision 24589 to svn://svn.ffmpeg.org/ffmpeg/trunk

7860b436 07/30/2010 08:42 AM Stefano Sabatini

Rename av_fill_image_linesizes() internal variables max_plane_step and
max_plane_step_comp by removing the "plane_" word, and add a comment
for explaining what they represent.

Increase readability.

Originally committed as revision 24588 to svn://svn.ffmpeg.org/ffmpeg/trunk

fc2db52e 07/30/2010 08:42 AM Stefano Sabatini

Make avpicture_fill() directly call av_fill_image_linesizes() and
av_fill_image_pointers() rather than their wrappers ff_fill_linesize()
and ff_fill_pointer().

Improve performance.

Originally committed as revision 24587 to svn://svn.ffmpeg.org/ffmpeg/trunk

7be5b730 07/30/2010 12:29 AM Stefano Sabatini

Make avfilter_default_get_video_buffer() use functions in
libavcore/imgutils.c rather than ff_fill_linesize() and
ff_fill_pointer().

Also remove a dependency on libavcodec.

Originally committed as revision 24586 to svn://svn.ffmpeg.org/ffmpeg/trunk

8fad2661 07/30/2010 12:05 AM Jean-Daniel Dupas

Make pkgconfig_generate() explicitly return 0 in instead of returning
without value when the target library is disabled.

If it does not explicitly return 0, when the last library is disabled
(swscale), the final "configure" exit value is 1, even if the
configure script is successfully executed. So it breaks scripts that...

3dec10cd 07/29/2010 11:58 PM Stefano Sabatini

Add APIchanges entry for the libavcore/imgutils.h addition.

Originally committed as revision 24584 to svn://svn.ffmpeg.org/ffmpeg/trunk

e7bd48a6 07/29/2010 11:44 PM Stefano Sabatini

Move fill_image_linesize() and fill_image_data_ptr() from
libavcodec/imgconvert.c and make them public in libavcore/imgutils.h,
with the names av_fill_image_linesizes() and av_fill_image_pointers().

Originally committed as revision 24583 to svn://svn.ffmpeg.org/ffmpeg/trunk

740dfe70 07/29/2010 10:45 PM Vitor Sessak

Fix compilation in x86_64. I broke it with r24580.

Originally committed as revision 24582 to svn://svn.ffmpeg.org/ffmpeg/trunk