OSDN Git Service

android-x86/external-ffmpeg.git
11 years agox86: cavs: Put mmx-specific code into its own init function
Diego Biurrun [Sat, 20 Apr 2013 20:03:19 +0000 (22:03 +0200)]
x86: cavs: Put mmx-specific code into its own init function

Before, this code was labeled as mmxext and enabled both for the
3dnow and the mmxext case.

11 years agox86: Remove some duplicate function declarations
Diego Biurrun [Sat, 20 Apr 2013 18:27:47 +0000 (20:27 +0200)]
x86: Remove some duplicate function declarations

11 years agoafifo: fix request_samples on the last frame in certain cases
Anton Khirnov [Tue, 16 Apr 2013 19:53:56 +0000 (21:53 +0200)]
afifo: fix request_samples on the last frame in certain cases

The current code can fail to return the last frame if it contains
exactly the requested number of samples.

Fixes the join filter test, which previously did not include the last
408 samples in most cases.

CC:libav-stable@libav.org

Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: Factor out WAVEFORMATEX parsing
Luca Barbato [Sat, 20 Apr 2013 11:36:48 +0000 (13:36 +0200)]
riff: Factor out WAVEFORMATEX parsing

Makes the code simpler to follow.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: Add ACTRAC3+ guid
Peter Ross [Sat, 20 Apr 2013 11:36:47 +0000 (13:36 +0200)]
riff: Add ACTRAC3+ guid

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: Perform full lookup on WAVEFORMATEXTENSIBLE subformat guid
Peter Ross [Sat, 20 Apr 2013 11:36:46 +0000 (13:36 +0200)]
riff: Perform full lookup on WAVEFORMATEXTENSIBLE subformat guid

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: Move guid structs and helper functions into riff
Peter Ross [Sat, 20 Apr 2013 11:36:45 +0000 (13:36 +0200)]
riff: Move guid structs and helper functions into riff

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: check for eof if chunk size and code are 0
Luca Barbato [Sat, 20 Apr 2013 11:36:44 +0000 (13:36 +0200)]
riff: check for eof if chunk size and code are 0

Prevent an infinite loop.

Inspired by a patch from Michael Niedermayer

CC: libav-stable@libav.org
Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoriff: K&R formatting cosmetics
Luca Barbato [Sat, 20 Apr 2013 11:36:43 +0000 (13:36 +0200)]
riff: K&R formatting cosmetics

Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoppc: hpeldsp: Include attributes.h
Martin Storsjö [Fri, 19 Apr 2013 22:29:17 +0000 (01:29 +0300)]
ppc: hpeldsp: Include attributes.h

This fixes building in configurations where altivec is disabled.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agox86: Remove unused inline asm instruction defines
Martin Storsjö [Fri, 19 Apr 2013 21:06:48 +0000 (00:06 +0300)]
x86: Remove unused inline asm instruction defines

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agovc1: Remove now unused variables
Martin Storsjö [Fri, 19 Apr 2013 21:04:20 +0000 (00:04 +0300)]
vc1: Remove now unused variables

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agocosmetics: bfin: Fix indentation in the dsputil init function
Martin Storsjö [Wed, 3 Apr 2013 11:11:38 +0000 (04:11 -0700)]
cosmetics: bfin: Fix indentation in the dsputil init function

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Remove non-8bpp draw_edge
Ronald S. Bultje [Thu, 14 Mar 2013 17:01:52 +0000 (10:01 -0700)]
dsputil: Remove non-8bpp draw_edge

It is never used.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Merge 9-10 bpp functions for get_pixels and draw_edge
Ronald S. Bultje [Tue, 26 Mar 2013 15:09:05 +0000 (17:09 +0200)]
dsputil: Merge 9-10 bpp functions for get_pixels and draw_edge

These only care about pixel storage unit size, not actual bits
used (i.e. they don't clip).

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Remove unused 32-bit functions
Ronald S. Bultje [Tue, 26 Mar 2013 15:07:22 +0000 (17:07 +0200)]
dsputil: Remove unused 32-bit functions

Previously, if dct_bits was set to 32, we used separate 32-bit
versions of these functions. Since dct_bits now is removed,
remove the unused 32-bit versions of the functions.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Remove dct_bits
Ronald S. Bultje [Tue, 26 Mar 2013 15:00:21 +0000 (17:00 +0200)]
dsputil: Remove dct_bits

dct_bits is never set except in h264, where it is never used,
thus remove it.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Remove hpel functions (moved to hpeldsp)
Ronald S. Bultje [Sun, 10 Mar 2013 21:23:46 +0000 (14:23 -0700)]
dsputil: Remove hpel functions (moved to hpeldsp)

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoalpha: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 23:53:07 +0000 (16:53 -0700)]
alpha: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agosparc: hpeldsp: Move vis half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 23:51:25 +0000 (16:51 -0700)]
sparc: hpeldsp: Move vis half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agosh4: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Michael Niedermayer [Sun, 10 Mar 2013 23:48:31 +0000 (16:48 -0700)]
sh4: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agobfin: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 23:40:23 +0000 (16:40 -0700)]
bfin: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoarm: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 23:16:45 +0000 (16:16 -0700)]
arm: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoppc: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 22:50:53 +0000 (15:50 -0700)]
ppc: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agox86: hpeldsp: Move half-pel assembly from dsputil to hpeldsp
Ronald S. Bultje [Sun, 10 Mar 2013 22:37:59 +0000 (15:37 -0700)]
x86: hpeldsp: Move half-pel assembly from dsputil to hpeldsp

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agomjpeg: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 21:23:52 +0000 (14:23 -0700)]
mjpeg: Use hpeldsp instead of dsputil for half-pel functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agosvq1enc: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 21:16:00 +0000 (14:16 -0700)]
svq1enc: Use hpeldsp instead of dsputil for half-pel functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agosvq3: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 21:12:58 +0000 (14:12 -0700)]
svq3: Use hpeldsp instead of dsputil for half-pel functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agompegvideo: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 20:55:07 +0000 (13:55 -0700)]
mpegvideo: Use hpeldsp instead of dsputil for half-pel functions

This also converts vc1, since that is mpegvideo-based.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agosvq1: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 20:38:45 +0000 (13:38 -0700)]
svq1: Use hpeldsp instead of dsputil for half-pel functions

This makes svq1 independent of dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agomimic: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 20:32:10 +0000 (13:32 -0700)]
mimic: Use hpeldsp instead of dsputil for half-pel functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agointerplayvideo: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 19:58:02 +0000 (12:58 -0700)]
interplayvideo: Use hpeldsp instead of dsputil for half-pel functions

This makes interplayvideo independent of dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agobink: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 16:12:01 +0000 (09:12 -0700)]
bink: Use hpeldsp instead of dsputil for half-pel functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoindeo3: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 15:19:27 +0000 (08:19 -0700)]
indeo3: Use hpeldsp instead of dsputil for half-pel functions

This makes the Indeo 3 decoder independent of dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agovp56: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 13:55:58 +0000 (06:55 -0700)]
vp56: Use hpeldsp instead of dsputil for half-pel functions

This makes vp5 and vp6 independent of dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agovp3: Use hpeldsp instead of dsputil for half-pel functions
Ronald S. Bultje [Sun, 10 Mar 2013 05:26:09 +0000 (21:26 -0800)]
vp3: Use hpeldsp instead of dsputil for half-pel functions

This makes vp3 independent of dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agohpeldsp: Add half-pel functions (currently copies of dsputil)
Ronald S. Bultje [Sat, 9 Mar 2013 23:14:35 +0000 (15:14 -0800)]
hpeldsp: Add half-pel functions (currently copies of dsputil)

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agofate: cosmetics: More sensible order for entries in video filter file
Diego Biurrun [Tue, 16 Apr 2013 14:46:40 +0000 (16:46 +0200)]
fate: cosmetics: More sensible order for entries in video filter file

11 years agofate: Rename video filters file and add separate video filters target
Diego Biurrun [Tue, 16 Apr 2013 14:43:49 +0000 (16:43 +0200)]
fate: Rename video filters file and add separate video filters target

11 years agofate: Split audio filters into their own separate file
Diego Biurrun [Tue, 16 Apr 2013 14:38:02 +0000 (16:38 +0200)]
fate: Split audio filters into their own separate file

11 years agofate: Reuse VREF and AREF variables where appropriate
Diego Biurrun [Tue, 16 Apr 2013 14:32:18 +0000 (16:32 +0200)]
fate: Reuse VREF and AREF variables where appropriate

11 years agoh264: fully support cropping.
Anton Khirnov [Mon, 18 Feb 2013 15:32:18 +0000 (16:32 +0100)]
h264: fully support cropping.

Based on a patch by Vittorio Giovara <vittorio.giovara@gmail.com>

Fixes Bug 378.

11 years agoff_get_buffer(): allocate the frame for max(coded,display) dimensions
Anton Khirnov [Mon, 18 Feb 2013 15:29:03 +0000 (16:29 +0100)]
ff_get_buffer(): allocate the frame for max(coded,display) dimensions

Needed e.g. for h264 cropping to work properly.

11 years agomp3dec: fallback to generic seeking when a TOC is not present
Michael Niedermayer [Thu, 20 Sep 2012 20:00:52 +0000 (22:00 +0200)]
mp3dec: fallback to generic seeking when a TOC is not present

Fixes seeking without a Xing/Info header.

CC: libav-stable@libav.org
Signed-off-by: Anton Khirnov <anton@khirnov.net>
11 years agoindeo3: check motion vectors.
Anton Khirnov [Tue, 16 Apr 2013 07:41:28 +0000 (09:41 +0200)]
indeo3: check motion vectors.

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agoindeo3: fix data size check
Anton Khirnov [Wed, 10 Apr 2013 07:59:36 +0000 (09:59 +0200)]
indeo3: fix data size check

The data offsets are relative to the bistream header, which is 16 bytes
after the start of the data.
Fixes invalid reads with corrupted files.

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agoindeo3: switch parsing the header to bytestream2
Anton Khirnov [Wed, 10 Apr 2013 07:40:20 +0000 (09:40 +0200)]
indeo3: switch parsing the header to bytestream2

Also add an additional sanity check to the alt_quant table.
Fixes invalid reads with corrupted files.

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agorv10: check that extradata is large enough
Anton Khirnov [Tue, 9 Apr 2013 18:33:25 +0000 (20:33 +0200)]
rv10: check that extradata is large enough

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agorv10: return meaningful error codes.
Anton Khirnov [Tue, 9 Apr 2013 18:11:09 +0000 (20:11 +0200)]
rv10: return meaningful error codes.

Also improve some error messages.

11 years agorv10: cosmetics, reformat
Anton Khirnov [Tue, 9 Apr 2013 18:06:44 +0000 (20:06 +0200)]
rv10: cosmetics, reformat

11 years agoqdm2: check that the FFT size is a power of 2
Anton Khirnov [Tue, 9 Apr 2013 13:25:20 +0000 (15:25 +0200)]
qdm2: check that the FFT size is a power of 2

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agosvq1dec: clip motion vectors to the frame size.
Anton Khirnov [Mon, 8 Apr 2013 20:15:54 +0000 (22:15 +0200)]
svq1dec: clip motion vectors to the frame size.

Fixes invalid reads for corrupted files.

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agosvq1dec: check that the reference frame has the same dimensions as the current one
Anton Khirnov [Mon, 8 Apr 2013 20:12:12 +0000 (22:12 +0200)]
svq1dec: check that the reference frame has the same dimensions as the current one

They can be different if the last keyframe failed to decode correctly.
Fixes possible invalid reads in such a case.

Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC:libav-stable@libav.org

11 years agoaf_channelmap: cosmetics, use an extra local variable to shorten the code
Anton Khirnov [Sun, 14 Apr 2013 10:10:50 +0000 (12:10 +0200)]
af_channelmap: cosmetics, use an extra local variable to shorten the code

11 years agoaf_channelmap: sanity check input channel indices in all cases.
Anton Khirnov [Sun, 14 Apr 2013 10:07:24 +0000 (12:07 +0200)]
af_channelmap: sanity check input channel indices in all cases.

Fixes invalid reads from non-existing channels.

CC:libav-stable@libav.org

11 years agocmdutils: Fix build with lavfi disabled
Anton Khirnov [Tue, 16 Apr 2013 14:09:16 +0000 (16:09 +0200)]
cmdutils: Fix build with lavfi disabled

Signed-off-by: Diego Biurrun <diego@biurrun.de>
11 years agoflvenc: do not mux more than one stream per type
Rafaël Carré [Mon, 15 Apr 2013 11:14:28 +0000 (13:14 +0200)]
flvenc: do not mux more than one stream per type

FLV does not support multiple audio or video streams.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
11 years agofate: Set the bitexact flag for vp3/vp5/vp6 tests
Martin Storsjö [Wed, 10 Apr 2013 08:48:57 +0000 (11:48 +0300)]
fate: Set the bitexact flag for vp3/vp5/vp6 tests

This is required since there are bit-inexact implementations
of the vp3 idct (for bfin).

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agovp3: Use full transpose for all IDCTs
Ronald S. Bultje [Tue, 12 Mar 2013 14:28:12 +0000 (07:28 -0700)]
vp3: Use full transpose for all IDCTs

This way, the special IDCT permutations are no longer needed. This
is similar to how H264 does it, and removes the dsputil dependency
imposed by the scantable code.

Also remove the unused type == 0 cases from the plain C version
of the idct.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agovp3: bfin: Transpose the IDCTs
Martin Storsjö [Wed, 3 Apr 2013 12:03:17 +0000 (05:03 -0700)]
vp3: bfin: Transpose the IDCTs

While this change isn't bitexact, the IDCTs weren't bitexact to
start with either.

This simplifies decoupling vp3 from dsputil.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agox86: Move constants to the only place where they are used
Ronald S. Bultje [Wed, 10 Apr 2013 09:34:10 +0000 (12:34 +0300)]
x86: Move constants to the only place where they are used

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agobfin: Match DEFUN_END macros to the right functions
Martin Storsjö [Wed, 10 Apr 2013 16:12:58 +0000 (19:12 +0300)]
bfin: Match DEFUN_END macros to the right functions

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agox86: dsputil: Move some ifdefs to avoid unused variable warnings
Diego Biurrun [Wed, 10 Apr 2013 23:31:15 +0000 (01:31 +0200)]
x86: dsputil: Move some ifdefs to avoid unused variable warnings

11 years agox86: dsputil: cosmetics: Remove two pointless variable indirections
Diego Biurrun [Wed, 10 Apr 2013 23:30:25 +0000 (01:30 +0200)]
x86: dsputil: cosmetics: Remove two pointless variable indirections

11 years agox86: dsputil: Refactor some ff_{avg|put}_pixels function declarations
Diego Biurrun [Thu, 11 Apr 2013 00:31:09 +0000 (02:31 +0200)]
x86: dsputil: Refactor some ff_{avg|put}_pixels function declarations

11 years agox86: dsputil: ff_h263_*_loop_filter declarations to a more suitable place
Diego Biurrun [Wed, 10 Apr 2013 23:19:44 +0000 (01:19 +0200)]
x86: dsputil: ff_h263_*_loop_filter declarations to a more suitable place

11 years agox86: h264qpel: int --> ptrdiff_t for some line_size parameters
Diego Biurrun [Wed, 10 Apr 2013 23:17:12 +0000 (01:17 +0200)]
x86: h264qpel: int --> ptrdiff_t for some line_size parameters

11 years agoFATE: add a test for the interlace filter
Vittorio Giovara [Thu, 4 Apr 2013 13:39:40 +0000 (15:39 +0200)]
FATE: add a test for the interlace filter

Signed-off-by: Anton Khirnov <anton@khirnov.net>
11 years agolavfi: new interlace filter
Vittorio Giovara [Tue, 12 Mar 2013 10:35:21 +0000 (11:35 +0100)]
lavfi: new interlace filter

Signed-off-by: Anton Khirnov <anton@khirnov.net>
11 years agocmdutils: allow -h filter=<name> to print information about a filter.
Anton Khirnov [Thu, 28 Mar 2013 08:29:00 +0000 (09:29 +0100)]
cmdutils: allow -h filter=<name> to print information about a filter.

11 years agoavtools: move media_type_string() from avprobe to cmdutils.
Anton Khirnov [Thu, 28 Mar 2013 08:10:09 +0000 (09:10 +0100)]
avtools: move media_type_string() from avprobe to cmdutils.

It will be useful outside of avprobe.

11 years agoavconv: print filter options in -h full output.
Anton Khirnov [Wed, 13 Mar 2013 08:10:34 +0000 (09:10 +0100)]
avconv: print filter options in -h full output.

11 years agoavconv: update to the new options syntax.
Anton Khirnov [Tue, 12 Mar 2013 18:31:08 +0000 (19:31 +0100)]
avconv: update to the new options syntax.

The separators were changed from ',' to '|' in all the filters.

11 years agolavfi: do not segfault on NULL passed to avfilter_get_by_name()
Anton Khirnov [Sun, 7 Apr 2013 05:33:24 +0000 (07:33 +0200)]
lavfi: do not segfault on NULL passed to avfilter_get_by_name()

11 years agolavfi: change the filter registering system to match the other libraries
Anton Khirnov [Wed, 3 Apr 2013 07:20:36 +0000 (09:20 +0200)]
lavfi: change the filter registering system to match the other libraries

Removes an arbitrary hardcoded limit on the number of filters.

11 years agolavfi: add a function for counting elements in AVFilterPad arrays.
Anton Khirnov [Thu, 28 Mar 2013 07:59:13 +0000 (08:59 +0100)]
lavfi: add a function for counting elements in AVFilterPad arrays.

The caller needs to know what valid indices can be passed to
avfilter_pad_get_name/type.

11 years agolavfi: mark filters with dynamic number of inputs or outputs with special flags
Anton Khirnov [Thu, 28 Mar 2013 07:51:36 +0000 (08:51 +0100)]
lavfi: mark filters with dynamic number of inputs or outputs with special flags

This will be useful in avtools in the following commits.
Any other caller might also want to know this information.

11 years agolavfi: add avfilter_init_dict() for initializing a filter with a dict.
Anton Khirnov [Sun, 17 Mar 2013 19:16:12 +0000 (20:16 +0100)]
lavfi: add avfilter_init_dict() for initializing a filter with a dict.

11 years agolavfi: add avfilter_init_str() to replace avfilter_init_filter().
Anton Khirnov [Sun, 17 Mar 2013 18:44:24 +0000 (19:44 +0100)]
lavfi: add avfilter_init_str() to replace avfilter_init_filter().

Drop the unused opaque parameter from its signature.

11 years agolavfi: make avfilter_free() remove the filter from its graph.
Anton Khirnov [Sun, 31 Mar 2013 11:02:55 +0000 (13:02 +0200)]
lavfi: make avfilter_free() remove the filter from its graph.

11 years agolavfi: add AVFilterContext.graph.
Anton Khirnov [Sun, 31 Mar 2013 10:17:07 +0000 (12:17 +0200)]
lavfi: add AVFilterContext.graph.

It will be useful in the following commits.

11 years agolavfi: deprecate avfilter_graph_add_filter().
Anton Khirnov [Sun, 31 Mar 2013 06:31:40 +0000 (08:31 +0200)]
lavfi: deprecate avfilter_graph_add_filter().

Since this function adds a standalone filter to a filter graph and we do
not support creating such filters, there is no reason for this function
to exist.

11 years agolavfi: replace avfilter_open() with avfilter_graph_alloc_filter().
Anton Khirnov [Sun, 31 Mar 2013 06:28:11 +0000 (08:28 +0200)]
lavfi: replace avfilter_open() with avfilter_graph_alloc_filter().

Since we do not support "standalone" filters not attached to an
AVFilterGraph, we should not have a public function to create such
filters. In addition that function is horribly named, the action it does
cannot be possibly described as "opening" a filter.

11 years agolavfi: merge avfiltergraph.h into avfilter.h
Anton Khirnov [Sat, 30 Mar 2013 21:31:18 +0000 (22:31 +0100)]
lavfi: merge avfiltergraph.h into avfilter.h

We do not support using filters without AVFilterGraph in practice
anyway, so there is no point in pretending we do.

11 years agolavfi: add const to the pads parameter of avfilter_pad_get_name/type
Anton Khirnov [Thu, 28 Mar 2013 08:31:42 +0000 (09:31 +0100)]
lavfi: add const to the pads parameter of avfilter_pad_get_name/type

11 years agolavfi: add const to AVFilterContext.filter.
Anton Khirnov [Sun, 31 Mar 2013 06:20:06 +0000 (08:20 +0200)]
lavfi: add const to AVFilterContext.filter.

lavfi should never modify the filter through that pointer.

11 years agodoc: document the prores encoder options
Anton Khirnov [Fri, 22 Mar 2013 23:13:44 +0000 (00:13 +0100)]
doc: document the prores encoder options

Adapted from the author's blag.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
11 years agoCosmetics, restore alphabetic order in atomics lists.
Anton Khirnov [Sat, 6 Apr 2013 17:04:56 +0000 (19:04 +0200)]
Cosmetics, restore alphabetic order in atomics lists.

11 years agoimg2enc: add an option for overwriting one file with subsequent images
Anton Khirnov [Thu, 4 Apr 2013 20:57:58 +0000 (22:57 +0200)]
img2enc: add an option for overwriting one file with subsequent images

Based on a patch by Michael Niedermayer <michaelni@gmx.at>.

11 years agoMove misplaced file author information where it belongs
Diego Biurrun [Wed, 10 Apr 2013 17:26:49 +0000 (19:26 +0200)]
Move misplaced file author information where it belongs

11 years agolavc: don't overwrite display dimensions with coded dimensions.
Anton Khirnov [Fri, 22 Mar 2013 06:59:10 +0000 (07:59 +0100)]
lavc: don't overwrite display dimensions with coded dimensions.

11 years agolavc: extend / update the [coded_]{width,height} doxy
Anton Khirnov [Fri, 22 Mar 2013 07:20:27 +0000 (08:20 +0100)]
lavc: extend / update the [coded_]{width,height} doxy

11 years agoh261: Move encoder/decoder shared table init to common code
Diego Biurrun [Fri, 5 Apr 2013 21:43:31 +0000 (23:43 +0200)]
h261: Move encoder/decoder shared table init to common code

11 years agoh261: Move shared data tables from a header to a proper C file
Diego Biurrun [Thu, 4 Apr 2013 12:56:01 +0000 (14:56 +0200)]
h261: Move shared data tables from a header to a proper C file

11 years agodsputil: Make dsputil selectable
Ronald S. Bultje [Tue, 12 Mar 2013 14:28:11 +0000 (07:28 -0700)]
dsputil: Make dsputil selectable

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoh264: Only initialize dsputil if error resilience is enabled
Ronald S. Bultje [Tue, 26 Mar 2013 14:43:57 +0000 (16:43 +0200)]
h264: Only initialize dsputil if error resilience is enabled

It is only used for error resilience. This allows building the
h264 decoder without dsputil, if error resilience is disabled.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoh264: Integrate clear_blocks calls with IDCT
Ronald S. Bultje [Tue, 19 Feb 2013 05:03:02 +0000 (21:03 -0800)]
h264: Integrate clear_blocks calls with IDCT

The non-intra-pcm branch in hl_decode_mb (simple, 8bpp) goes from 700
to 672 cycles, and the complete loop of decode_mb_cabac and hl_decode_mb
(in the decode_slice loop) goes from 1759 to 1733 cycles on the clip
tested (cathedral), i.e. almost 30 cycles per mb faster.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoh264: Clear the mb members via memset instead of using dsputil
Martin Storsjö [Wed, 3 Apr 2013 14:23:01 +0000 (17:23 +0300)]
h264: Clear the mb members via memset instead of using dsputil

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agodsputil: Make square put/avg_pixels functions local to h264qpel
Ronald S. Bultje [Sat, 9 Mar 2013 04:16:08 +0000 (20:16 -0800)]
dsputil: Make square put/avg_pixels functions local to h264qpel

Put a copy of the 8bit functions only in dsputil, where they are
used for some other things (e.g. mpeg4qpel, mspel, cavsqpel).

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agox86inc: Fix number of operands for cmp* instructions
Christophe Gisquet [Sun, 7 Apr 2013 19:29:52 +0000 (19:29 +0000)]
x86inc: Fix number of operands for cmp* instructions

cmp{p,s}{s,d} instructions do take an imm8 operand.

Signed-off-by: Diego Biurrun <diego@biurrun.de>