platform/upstream/libav.git
12 years agoavprobe: close opened codecs after use
Janne Grunau [Mon, 14 May 2012 10:23:10 +0000 (12:23 +0200)]
avprobe: close opened codecs after use

Fixes "memleak" on closing avprobe to make valgrind happy.

12 years agovcr1: Add vcr1_ prefixes to all static functions with generic names.
Diego Biurrun [Mon, 14 May 2012 10:36:45 +0000 (12:36 +0200)]
vcr1: Add vcr1_ prefixes to all static functions with generic names.

12 years agovcr1: Fix return type of common_init to match the function pointer signature.
Diego Biurrun [Mon, 14 May 2012 10:33:04 +0000 (12:33 +0200)]
vcr1: Fix return type of common_init to match the function pointer signature.

libavcodec/vcr1.c:182: warning: initialization from incompatible pointer type

12 years agovcr1enc: Replace obsolete get_bit_count by put_bits_count/flush_put_bits.
Diego Biurrun [Fri, 13 Apr 2012 06:52:08 +0000 (08:52 +0200)]
vcr1enc: Replace obsolete get_bit_count by put_bits_count/flush_put_bits.

12 years agomotion-test: remove disabled code
Diego Biurrun [Sun, 17 Jul 2011 11:30:24 +0000 (13:30 +0200)]
motion-test: remove disabled code

12 years agogxfenc: remove disabled half-implemented MJPEG tag
Diego Biurrun [Fri, 29 Jul 2011 10:02:20 +0000 (12:02 +0200)]
gxfenc: remove disabled half-implemented MJPEG tag

12 years agox86: use more standard construct for setting ASM functions in FFT code
Vitor Sessak [Sat, 12 May 2012 06:59:00 +0000 (08:59 +0200)]
x86: use more standard construct for setting ASM functions in FFT code

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agofate: westwood-aud: disable decoding
Mans Rullgard [Sat, 12 May 2012 12:17:38 +0000 (13:17 +0100)]
fate: westwood-aud: disable decoding

The codec (adpcm-ima-ws) is tested elsewhere.  Using framecrc output
provides more information than a single md5 if something goes wrong.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: caf: disable decoding
Mans Rullgard [Sat, 12 May 2012 11:39:59 +0000 (12:39 +0100)]
fate: caf: disable decoding

This is intended as a demuxer test and the file contains pcm_s16be
audio which is tested elsewhere.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: film-cvid: drop pcm audio and rename test
Mans Rullgard [Fri, 11 May 2012 19:59:58 +0000 (20:59 +0100)]
fate: film-cvid: drop pcm audio and rename test

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: d-cinema-demux: drop unnecessary flags
Mans Rullgard [Fri, 11 May 2012 17:30:44 +0000 (18:30 +0100)]
fate: d-cinema-demux: drop unnecessary flags

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: split off dpcm-interplay from interplay-mve tests
Mans Rullgard [Fri, 11 May 2012 15:48:17 +0000 (16:48 +0100)]
fate: split off dpcm-interplay from interplay-mve tests

These two files use the same audio codec so only one test for
this is needed.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: rename funcom-iss to adpcm-ima-iss
Mans Rullgard [Fri, 11 May 2012 16:43:54 +0000 (17:43 +0100)]
fate: rename funcom-iss to adpcm-ima-iss

This matches the name of the codec.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: rename cryo-apc to adpcm-ima-apc
Mans Rullgard [Fri, 11 May 2012 16:38:41 +0000 (17:38 +0100)]
fate: rename cryo-apc to adpcm-ima-apc

This matches the name of the tested codec.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: rename adpcm-psx-str-v3 to adpcm-xa
Mans Rullgard [Fri, 11 May 2012 15:25:34 +0000 (16:25 +0100)]
fate: rename adpcm-psx-str-v3 to adpcm-xa

This matches the name of the codec tested.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: split off adpcm-ms-mono test from dxa-feeble
Mans Rullgard [Fri, 11 May 2012 16:08:29 +0000 (17:08 +0100)]
fate: split off adpcm-ms-mono test from dxa-feeble

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: split off adpcm-ima-ws test from vqa-cc
Mans Rullgard [Fri, 11 May 2012 14:36:57 +0000 (15:36 +0100)]
fate: split off adpcm-ima-ws test from vqa-cc

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: add adpcm-ima-smjpeg test
Mans Rullgard [Fri, 11 May 2012 14:16:43 +0000 (15:16 +0100)]
fate: add adpcm-ima-smjpeg test

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: split off adpcm-ima-amv from amv test
Mans Rullgard [Fri, 11 May 2012 12:33:35 +0000 (13:33 +0100)]
fate: split off adpcm-ima-amv from amv test

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: separate bmv audio and video tests
Mans Rullgard [Fri, 11 May 2012 13:58:31 +0000 (14:58 +0100)]
fate: separate bmv audio and video tests

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: separate delphine-cin audio and video tests
Mans Rullgard [Fri, 11 May 2012 03:11:43 +0000 (04:11 +0100)]
fate: separate delphine-cin audio and video tests

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: truemotion1: disable audio
Mans Rullgard [Fri, 11 May 2012 16:16:58 +0000 (17:16 +0100)]
fate: truemotion1: disable audio

These tests include adpcm-ima-dk3 audio which is tested elsewhere.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: qtrle: disable audio in all tests
Mans Rullgard [Fri, 11 May 2012 15:30:00 +0000 (16:30 +0100)]
fate: qtrle: disable audio in all tests

These files contain mace6 audio which is tested elsewhere.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: pcm-planar: disable video
Mans Rullgard [Fri, 11 May 2012 12:49:13 +0000 (13:49 +0100)]
fate: pcm-planar: disable video

This file has eamad video which is tested elsewhere.
Also rename the test to reflect this change.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: mtv: disable video decoding
Mans Rullgard [Fri, 11 May 2012 14:47:37 +0000 (15:47 +0100)]
fate: mtv: disable video decoding

This test contains raw rgb565le video.  Converting to rgb24 serves
no useful purpose here.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agodoc: misc improvements for the Windows section
Diego Biurrun [Thu, 10 May 2012 12:18:10 +0000 (14:18 +0200)]
doc: misc improvements for the Windows section

Fix some orthography, wording and grammar issues; update the SDL section
with more current instructions; simplify lib.exe example command line;
drop outdated comments about libnut.

12 years agodoc: fix link to fate.libav.org
Diego Elio Pettenò [Sat, 12 May 2012 23:22:06 +0000 (16:22 -0700)]
doc: fix link to fate.libav.org

Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.eu>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
12 years agolavfi: autoinsert resample filter when necessary.
Anton Khirnov [Fri, 4 May 2012 13:35:12 +0000 (15:35 +0200)]
lavfi: autoinsert resample filter when necessary.

12 years agolavfi: add lavr-based audio resampling filter.
Anton Khirnov [Fri, 4 May 2012 13:09:17 +0000 (15:09 +0200)]
lavfi: add lavr-based audio resampling filter.

12 years agox86: vc1: drop MMX loop filter implementation, which uses MMX2 instructions.
Michael Kostylev [Fri, 11 May 2012 10:26:44 +0000 (12:26 +0200)]
x86: vc1: drop MMX loop filter implementation, which uses MMX2 instructions.

12 years agofate: employ better names and add a convenient shorthand for vp6 tests
Diego Biurrun [Wed, 9 May 2012 02:02:01 +0000 (04:02 +0200)]
fate: employ better names and add a convenient shorthand for vp6 tests

12 years agoarm/neon: dsputil: use correct size specifiers on vld1/vst1
Mans Rullgard [Thu, 10 May 2012 16:40:30 +0000 (17:40 +0100)]
arm/neon: dsputil: use correct size specifiers on vld1/vst1

Change the size specifiers to match the actual element sizes
of the data.  This makes no practical difference with strict
alignment checking disabled (the default) other than somewhat
documenting the code.  With strict alignment checking on, it
avoids trapping the unaligned loads.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoarm: dsputil: prettify some conditional instructions in put_pixels macros
Mans Rullgard [Thu, 10 May 2012 15:24:33 +0000 (16:24 +0100)]
arm: dsputil: prettify some conditional instructions in put_pixels macros

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agovqavideo: change x/y loop counters to the usual pattern
Mans Rullgard [Mon, 23 Apr 2012 11:51:28 +0000 (12:51 +0100)]
vqavideo: change x/y loop counters to the usual pattern

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoavconv: use lrint() for rounding double timestamps
Mans Rullgard [Thu, 10 May 2012 11:19:10 +0000 (12:19 +0100)]
avconv: use lrint() for rounding double timestamps

Converting the double to float for lrintf() loses precision when
the value is not exactly representable as a single-precision float.
Apart from being inaccurate, this causes discrepancies in some
configurations due to differences in rounding.

Note that the changed timestamp in the vc1-ism test is a bogus,
made-up value.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agorv40dsp x86: MMX/MMX2/3DNow/SSE2/SSSE3 implementations of MC
Christophe Gisquet [Thu, 19 Apr 2012 20:36:17 +0000 (22:36 +0200)]
rv40dsp x86: MMX/MMX2/3DNow/SSE2/SSSE3 implementations of MC

Code mostly inspired by vp8's MC, however:
- its MMX2 horizontal filter is worse because it can't take advantage of
  the coefficient redundancy
- that same coefficient redundancy allows better code for non-SSSE3 versions

Benchmark (rounded to tens of unit):
        V8x8  H8x8  2D8x8  V16x16  H16x16  2D16x16
C       445    358   985    1785    1559    3280
MMX*    219    271   478     714     929    1443
SSE2    131    158   294     425     515     892
SSSE3   120    122   248     387     390     763

End result is overall around a 15% speedup for SSSE3 version (on 6 sequences);
all loop filter functions now take around 55% of decoding time, while luma MC
dsp functions are around 6%, chroma ones are 1.3% and biweight around 2.3%.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agoape: Use unsigned integer maths
Christophe Gisquet [Tue, 28 Feb 2012 16:42:12 +0000 (17:42 +0100)]
ape: Use unsigned integer maths

This involves a division that should be a shift.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agoarm: dsputil: fix overreads in put/avg_pixels functions
Mans Rullgard [Wed, 9 May 2012 23:55:18 +0000 (00:55 +0100)]
arm: dsputil: fix overreads in put/avg_pixels functions

The vertically interpolating variants of these functions read
ahead one line to optimise the loop.  On the last line processed,
this might be outside the buffer.  Fix these invalid reads by
processing the last line outside the loop.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoh264: K&R formatting cosmetics for header files (part II/II)
Diego Biurrun [Mon, 7 May 2012 12:13:57 +0000 (14:13 +0200)]
h264: K&R formatting cosmetics for header files (part II/II)

12 years agoh264: K&R formatting cosmetics for header files (part I/II)
Diego Biurrun [Mon, 7 May 2012 12:13:23 +0000 (14:13 +0200)]
h264: K&R formatting cosmetics for header files (part I/II)

12 years agortmp: Implement check bandwidth notification.
Samuel Pitoiset [Tue, 8 May 2012 22:58:09 +0000 (00:58 +0200)]
rtmp: Implement check bandwidth notification.

According to the behaviour of librtmp, it is recommended to send this
message to the server after receiving the 'onBWDone' callback in order
to do bandwidth checking and improve compatibility with some servers.

12 years agortmp: Support 'rtmp_swfurl', an option which specifies the URL of the SWF player.
Samuel Pitoiset [Wed, 9 May 2012 00:12:16 +0000 (02:12 +0200)]
rtmp: Support 'rtmp_swfurl', an option which specifies the URL of the SWF player.

12 years agortmp: Support 'rtmp_flashver', an option which overrides the version of the Flash...
Samuel Pitoiset [Wed, 9 May 2012 00:12:15 +0000 (02:12 +0200)]
rtmp: Support 'rtmp_flashver', an option which overrides the version of the Flash plugin.

12 years agortmp: Support 'rtmp_tcurl', an option which overrides the URL of the target stream.
Samuel Pitoiset [Wed, 9 May 2012 00:12:14 +0000 (02:12 +0200)]
rtmp: Support 'rtmp_tcurl', an option which overrides the URL of the target stream.

Signed-off-by: Martin Storsjö <martin@martin.st>
12 years agocmdutils: Add fallback case to switch in check_stream_specifier().
Diego Biurrun [Wed, 11 Apr 2012 08:31:02 +0000 (10:31 +0200)]
cmdutils: Add fallback case to switch in check_stream_specifier().

This fixes the warning:
cmdutils.c:897: warning: ‘type’ may be used uninitialized in this function

12 years agosctp: be consistent with socket option level
Sean McGovern [Wed, 9 May 2012 06:13:16 +0000 (02:13 -0400)]
sctp: be consistent with socket option level

Replace SOL_SCTP by the more portable IPPROTO_SCTP.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agoconfigure: Add _XOPEN_SOURCE=600 to Solaris preprocessor flags.
Sean McGovern [Wed, 9 May 2012 06:13:15 +0000 (02:13 -0400)]
configure: Add _XOPEN_SOURCE=600 to Solaris preprocessor flags.

This is needed to expose some networking APIs.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agovcr1enc: drop pointless empty encode_init() wrapper function
Diego Biurrun [Thu, 12 Apr 2012 16:55:25 +0000 (18:55 +0200)]
vcr1enc: drop pointless empty encode_init() wrapper function

12 years agovcr1: drop pointless write-only AVCodecContext member from VCR1Context
Diego Biurrun [Thu, 12 Apr 2012 16:52:16 +0000 (18:52 +0200)]
vcr1: drop pointless write-only AVCodecContext member from VCR1Context

12 years agovcr1: group encoder code together to save #ifdefs
Diego Biurrun [Mon, 9 Apr 2012 16:11:35 +0000 (18:11 +0200)]
vcr1: group encoder code together to save #ifdefs

12 years agovcr1: cosmetics: K&R prettyprinting, typos, parentheses, dead code, comments
Diego Biurrun [Sun, 8 Apr 2012 11:05:17 +0000 (13:05 +0200)]
vcr1: cosmetics: K&R prettyprinting, typos, parentheses, dead code, comments

12 years agomov: make one comment slightly more specific
Diego Biurrun [Thu, 23 Feb 2012 10:34:28 +0000 (11:34 +0100)]
mov: make one comment slightly more specific

12 years agolavr: replace the SSE version of ff_conv_fltp_to_flt_6ch() with SSE4 and AVX
Justin Ruggles [Thu, 3 May 2012 19:23:32 +0000 (15:23 -0400)]
lavr: replace the SSE version of ff_conv_fltp_to_flt_6ch() with SSE4 and AVX

The current SSE version is slower than the MMX version on Athlon64 and Sandy
Bridge, but the SSE4 and AVX versions are faster on Sandy Bridge.

12 years agolavfi: move audio-related functions to a separate file.
Anton Khirnov [Mon, 7 May 2012 09:21:38 +0000 (11:21 +0200)]
lavfi: move audio-related functions to a separate file.

This is easier to follow than having them randomly scattered in
avfilter.c and defaults.c.

12 years agolavfi: remove some audio-related function from public API.
Anton Khirnov [Mon, 7 May 2012 08:51:23 +0000 (10:51 +0200)]
lavfi: remove some audio-related function from public API.

Those functions are only useful inside filters. It is better to not
support user filters until the API is more stable.

This breaks audio filtering API and ABI in theory, but since it's
unusable right now this shouldn't be a problem.

12 years agolavfi: change AVFilterLink.sample_rate from int64_t to int on next bump
Anton Khirnov [Wed, 9 May 2012 06:43:07 +0000 (08:43 +0200)]
lavfi: change AVFilterLink.sample_rate from int64_t to int on next bump

There is no real reason for it to be 64bit, it's just a plain int in the
rest of Libav.

12 years agolavfi: change AVFilterBufferRefAudioProps.sample_rate from uint32_t to int
Anton Khirnov [Sun, 6 May 2012 08:40:11 +0000 (10:40 +0200)]
lavfi: change AVFilterBufferRefAudioProps.sample_rate from uint32_t to int

There's no reason for it to be explicitly 32 bits. It's declared as a
plain int in all other places in Libav.

This breaks audio filtering API and ABI in theory, but since it's
unusable right now this shouldn't be a problem.

12 years agolavfi: simplify signature for avfilter_get_audio_buffer() and friends
Stefano Sabatini [Tue, 30 Aug 2011 21:22:29 +0000 (23:22 +0200)]
lavfi: simplify signature for avfilter_get_audio_buffer() and friends

The additional parameters are just complicating the function interface.

Assume that a requested samples buffer will *always* have the format
specified in the requested link.

This breaks audio filtering API and ABI in theory, but since it's
unusable right now this shouldn't be a problem.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agolavfi: use avfilter_get_audio_buffer_ref_from_arrays() in avfilter_default_get_audio_...
Stefano Sabatini [Sat, 15 Jan 2011 17:48:37 +0000 (18:48 +0100)]
lavfi: use avfilter_get_audio_buffer_ref_from_arrays() in avfilter_default_get_audio_buffer

12 years agolavfi: cleanup avfilter_get_audio_buffer() and pals.
Stefano Sabatini [Tue, 1 Feb 2011 11:34:23 +0000 (12:34 +0100)]
lavfi: cleanup avfilter_get_audio_buffer() and pals.

Remove AVFilterBufferRefAudioProps.size, and use nb_samples in its place
everywhere.
This is required as the size in the audio buffer may be aligned, so it
may not contain a well defined number of samples.

Also remove the useless planar parameter, which can be deduced from the
sample format.

This is technically an API and ABI break, but since the audio part of
lavfi is not usable now, this should not be a problem in practice.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agolibxvid: Give more suitable names to libxvid-related files.
Diego Biurrun [Wed, 9 May 2012 15:20:36 +0000 (17:20 +0200)]
libxvid: Give more suitable names to libxvid-related files.

12 years agolibxvid: Separate libxvid encoder from libxvid rate control code.
Diego Biurrun [Tue, 10 Apr 2012 23:11:08 +0000 (01:11 +0200)]
libxvid: Separate libxvid encoder from libxvid rate control code.

This allows compiling the Xvid rate control code without the encoder.

12 years agojpeglsdec: Remove write-only variable in ff_jpegls_decode_lse().
Diego Biurrun [Thu, 19 Apr 2012 12:55:49 +0000 (14:55 +0200)]
jpeglsdec: Remove write-only variable in ff_jpegls_decode_lse().

libavcodec/jpeglsdec.c:54:9: warning: variable ‘len’ set but not used

12 years agofate: cosmetics: lowercase some comments
Diego Biurrun [Thu, 26 Apr 2012 13:04:20 +0000 (15:04 +0200)]
fate: cosmetics: lowercase some comments

12 years agofate: Give more consistent names to some RealVideo/RealAudio tests.
Diego Biurrun [Thu, 26 Apr 2012 12:55:37 +0000 (14:55 +0200)]
fate: Give more consistent names to some RealVideo/RealAudio tests.

12 years agolavfi: add avfilter_get_audio_buffer_ref_from_arrays().
Stefano Sabatini [Sun, 30 Jan 2011 23:07:41 +0000 (00:07 +0100)]
lavfi: add avfilter_get_audio_buffer_ref_from_arrays().

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agolavfi: add extended_data to AVFilterBuffer.
Anton Khirnov [Mon, 16 Apr 2012 12:01:26 +0000 (14:01 +0200)]
lavfi: add extended_data to AVFilterBuffer.

This is similar to what has previously been done in AVFrame to allow
dealing with more than 8 channels.

12 years agolavc: check that extended_data is properly set in avcodec_encode_audio2().
Anton Khirnov [Mon, 7 May 2012 14:56:20 +0000 (16:56 +0200)]
lavc: check that extended_data is properly set in avcodec_encode_audio2().

12 years agolavc: pad last audio frame with silence when needed.
Anton Khirnov [Mon, 7 May 2012 11:55:03 +0000 (13:55 +0200)]
lavc: pad last audio frame with silence when needed.

12 years agosamplefmt: add a function for filling a buffer with silence.
Anton Khirnov [Tue, 8 May 2012 11:56:37 +0000 (13:56 +0200)]
samplefmt: add a function for filling a buffer with silence.

12 years agosamplefmt: add a function for copying audio samples.
Anton Khirnov [Sun, 6 May 2012 12:10:38 +0000 (14:10 +0200)]
samplefmt: add a function for copying audio samples.

12 years agolavr: do not try to copy to uninitialized output audio data.
Anton Khirnov [Wed, 9 May 2012 04:36:29 +0000 (06:36 +0200)]
lavr: do not try to copy to uninitialized output audio data.

This would happen at least when lavr is used as a fifo with no
conversion.

12 years agolavr: make avresample_read() with NULL output discard samples.
Anton Khirnov [Wed, 9 May 2012 04:35:13 +0000 (06:35 +0200)]
lavr: make avresample_read() with NULL output discard samples.

12 years agofate: split idroq audio and video into separate tests
Mans Rullgard [Tue, 8 May 2012 23:59:40 +0000 (00:59 +0100)]
fate: split idroq audio and video into separate tests

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: improve dependencies
Mans Rullgard [Tue, 8 May 2012 21:07:26 +0000 (22:07 +0100)]
fate: improve dependencies

This makes only tests actually using avconv depend on it.
The remaining tests already depend on what they need.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: add convenient shorthands for ea-vp6, libavcodec, libavutil tests
Diego Biurrun [Thu, 26 Apr 2012 12:05:57 +0000 (12:05 +0000)]
fate: add convenient shorthands for ea-vp6, libavcodec, libavutil tests

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: split some combined tests into separate audio and video tests
Diego Biurrun [Fri, 27 Apr 2012 14:17:42 +0000 (14:17 +0000)]
fate: split some combined tests into separate audio and video tests

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: fix dependencies for probe tests
Mans Rullgard [Tue, 8 May 2012 10:29:09 +0000 (11:29 +0100)]
fate: fix dependencies for probe tests

Only the probe tests should depend on avprobe and these should
be enabled only if avprobe is configured.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agomips: intreadwrite: fix inline asm for gcc 4.8
Mans Rullgard [Sat, 5 May 2012 22:38:15 +0000 (23:38 +0100)]
mips: intreadwrite: fix inline asm for gcc 4.8

Just like gcc 4.6 and later on ARM, gcc 4.8 on MIPS generates
inefficient code when a known-unaligned location is used as a
memory input operand.  This applies the same fix as has been
previously done to the ARM version of the code.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agomips: intreadwrite: remove unnecessary inline asm
Mans Rullgard [Sat, 5 May 2012 22:24:51 +0000 (23:24 +0100)]
mips: intreadwrite: remove unnecessary inline asm

GCC actually handles unaligned accesses correctly in all cases
except, absurdly, 32-bit loads on mips64.  The remaining asm is
thus not needed, and removing it results in better code.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agolavfi: add video buffer sink, and use it in avtools
Anton Khirnov [Fri, 27 Apr 2012 04:56:56 +0000 (06:56 +0200)]
lavfi: add video buffer sink, and use it in avtools

Also add the public interface libavfilter/buffersink.h.

Based on a commit by Stefano Sabatini.

12 years agolavfi: add a function for copying properties from AVFilterBufferRef->AVFrame
Anton Khirnov [Fri, 27 Apr 2012 15:27:40 +0000 (17:27 +0200)]
lavfi: add a function for copying properties from AVFilterBufferRef->AVFrame

Based on a commit by Stefano Sabatini <stefano.sabatini-lala@poste.it>

12 years agosctp: Initial tcp-alike sctp support with streams
Luca Barbato [Sat, 21 Apr 2012 12:03:05 +0000 (14:03 +0200)]
sctp: Initial tcp-alike sctp support with streams

Signed-off-by: Jordi Ortiz <nenjordi@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
12 years agolibschroedingerdec: Change AVPicture to AVFrame and use SchroTag to store pts
Jordi Ortiz [Tue, 8 May 2012 17:20:32 +0000 (19:20 +0200)]
libschroedingerdec: Change AVPicture to AVFrame and use SchroTag to store pts

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
12 years agooptions_table: Add some missing #includes to fix "make checkheaders".
Alex Converse [Mon, 7 May 2012 23:25:12 +0000 (16:25 -0700)]
options_table: Add some missing #includes to fix "make checkheaders".

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agodoc: Replace a stray reference to the old '-intra' flag.
Alex Converse [Mon, 7 May 2012 18:57:11 +0000 (11:57 -0700)]
doc: Replace a stray reference to the old '-intra' flag.

12 years agortsp: Add content-type message header parsing
Jordi Ortiz [Fri, 4 May 2012 15:50:31 +0000 (17:50 +0200)]
rtsp: Add content-type message header parsing

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
12 years agortmp: Support 'rtmp_live', an option which specifies if the media is a live stream.
Samuel Pitoiset [Sat, 5 May 2012 17:33:26 +0000 (19:33 +0200)]
rtmp: Support 'rtmp_live', an option which specifies if the media is a live stream.

Signed-off-by: Martin Storsjö <martin@martin.st>
12 years agoav_samples_fill_array: Mark unmodified function argument as const.
Diego Biurrun [Wed, 11 Apr 2012 12:51:57 +0000 (14:51 +0200)]
av_samples_fill_array: Mark unmodified function argument as const.

libavcodec/utils.c:274: warning: passing argument 3 of ‘av_samples_fill_arrays’ discards qualifiers from pointer target type
./libavutil/samplefmt.h:151: note: expected ‘uint8_t *’ but argument is of type ‘const uint8_t *’

12 years agolagarith: add YUY2 decoding support
Kostya Shishkov [Mon, 7 May 2012 17:43:52 +0000 (19:43 +0200)]
lagarith: add YUY2 decoding support

Unlike other variants, for YUY2 we need to use different prediction:
* on line 0 for luma we should left predict starting from the second pixel
* on line 1 we should left predict first 4 pixels for luma and 2 for chroma
* median prediction employed here is taken directly from HuffYUV

12 years agoSupport decoding unaligned rgb24 lagarith.
Carl Eugen Hoyos [Tue, 8 May 2012 04:25:53 +0000 (22:25 -0600)]
Support decoding unaligned rgb24 lagarith.

Signed-off-by: Kostya Shishkov <kostya.shishkov@gmail.com>
12 years agodv: Split profile handling code into a separate file.
Diego Biurrun [Wed, 18 Apr 2012 15:55:49 +0000 (17:55 +0200)]
dv: Split profile handling code into a separate file.

12 years agoflvenc: use AVFormatContext, not AVCodecContext for logging.
Anton Khirnov [Mon, 7 May 2012 14:09:30 +0000 (16:09 +0200)]
flvenc: use AVFormatContext, not AVCodecContext for logging.

Encoder tag being used for muxer messages is confusing.

12 years agomov: Remove write-only variable in mov_read_chan().
Diego Biurrun [Thu, 19 Apr 2012 12:53:23 +0000 (14:53 +0200)]
mov: Remove write-only variable in mov_read_chan().

libavformat/mov.c:597:25: warning: variable ‘cflags’ set but not used

12 years agofate: Change the probe-format refs to match the final text format committed.
Alex Converse [Mon, 7 May 2012 18:01:38 +0000 (11:01 -0700)]
fate: Change the probe-format refs to match the final text format committed.

12 years agofate: Add avprobe as a make dependency
Alex Converse [Mon, 7 May 2012 18:00:54 +0000 (11:00 -0700)]
fate: Add avprobe as a make dependency

12 years agoAdd probe fate tests to test for regressions in detecting media types.
Ronald S. Bultje [Thu, 3 May 2012 18:01:00 +0000 (11:01 -0700)]
Add probe fate tests to test for regressions in detecting media types.

Signed-off-by: Alex Converse <alex.converse@gmail.com>
12 years agofate: Add oneline comparison method
Mans Rullgard [Thu, 3 May 2012 22:43:11 +0000 (15:43 -0700)]
fate: Add oneline comparison method

Signed-off-by: Alex Converse <alex.converse@gmail.com>
12 years agoqdm2: clip array indices returned by qdm2_get_vlc().
Ronald S. Bultje [Wed, 2 May 2012 16:12:46 +0000 (16:12 +0000)]
qdm2: clip array indices returned by qdm2_get_vlc().

Prevents subsequent overreads when these numbers are used as indices
in arrays.

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC: libav-stable@libav.org
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
12 years agoavplay: properly close/reopen AVAudioResampleContext on channel layout change
Justin Ruggles [Thu, 3 May 2012 22:57:02 +0000 (18:57 -0400)]
avplay: properly close/reopen AVAudioResampleContext on channel layout change

fixes Bug#280