ffmpeg.git
6 years agolibopencore-amrwb: Make AMR-WB ifdeffery more precise
Diego Biurrun [Sat, 16 Feb 2013 22:05:04 +0000 (23:05 +0100)]
libopencore-amrwb: Make AMR-WB ifdeffery more precise

The library might provide an encoder in the future, so it's better to
check for the presence of the decoder rather than just the library.

CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
6 years agolibopencore-amr: Conditionally compile decoder and encoder bits
Diego Biurrun [Sat, 16 Feb 2013 22:05:03 +0000 (23:05 +0100)]
libopencore-amr: Conditionally compile decoder and encoder bits

CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
6 years agolibopencore-amrnb: cosmetics: Group all encoder-related code together
Diego Biurrun [Sat, 16 Feb 2013 22:05:02 +0000 (23:05 +0100)]
libopencore-amrnb: cosmetics: Group all encoder-related code together

CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
6 years agodsputil: convert remaining op_pixels_func
Luca Barbato [Sat, 9 Feb 2013 23:17:20 +0000 (00:17 +0100)]
dsputil: convert remaining op_pixels_func

Convert to diffptr_t the line_size parameters still int.

Remove all the warnings in dsputil.c

6 years agoroqvideodec: check dimensions validity
Michael Niedermayer [Thu, 29 Nov 2012 14:18:17 +0000 (15:18 +0100)]
roqvideodec: check dimensions validity

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit 3ae610451170cd5a28b33950006ff0bd23036845)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit fee26d352a52eb9f7fcd8d9167fb4a5ba015b612)

CC: libav-stable@libav.org
Signed-off-by: Reinhard Tartler <siretart@tauware.de>
6 years agovqavideo: check chunk sizes before reading chunks
Michael Niedermayer [Fri, 25 Jan 2013 05:11:59 +0000 (06:11 +0100)]
vqavideo: check chunk sizes before reading chunks

Fixes out of array writes

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit ab6c9332bfa1e20127a16392a0b85a4aa4840889)

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit 13093f9767b922661132a3c1f4b5ba2c7338b660)

CC: libav-stable@libav.org
Signed-off-by: Reinhard Tartler <siretart@tauware.de>
6 years agoqdm2: check array index before use, fix out of array accesses
Michael Niedermayer [Fri, 30 Nov 2012 22:59:40 +0000 (23:59 +0100)]
qdm2: check array index before use, fix out of array accesses

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit a7ee6281f7ef1c29284e3a4cadfe0f227ffde1ed)

CC: libav-stable@libav.org
Signed-off-by: Reinhard Tartler <siretart@tauware.de>
6 years agoerror_resilience: add required headers.
Anton Khirnov [Sat, 16 Feb 2013 19:58:18 +0000 (20:58 +0100)]
error_resilience: add required headers.

6 years agoh264: deMpegEncContextize
Anton Khirnov [Sun, 3 Feb 2013 10:10:05 +0000 (11:10 +0100)]
h264: deMpegEncContextize

Most of the changes are just trivial are just trivial replacements of
fields from MpegEncContext with equivalent fields in H264Context.
Everything in h264* other than h264.c are those trivial changes.

The nontrivial parts are:
1) extracting a simplified version of the frame management code from
   mpegvideo.c. We don't need last/next_picture anymore, since h264 uses
   its own more complex system already and those were set only to appease
   the mpegvideo parts.
2) some tables that need to be allocated/freed in appropriate places.
3) hwaccels -- mostly trivial replacements.
   for dxva, the draw_horiz_band() call is moved from
   ff_dxva2_common_end_frame() to per-codec end_frame() callbacks,
   because it's now different for h264 and MpegEncContext-based
   decoders.
4) svq3 -- it does not use h264 complex reference system, so I just
   added some very simplistic frame management instead and dropped the
   use of ff_h264_frame_start(). Because of this I also had to move some
   initialization code to svq3.

Additional fixes for chroma format and bit depth changes by
Janne Grunau <janne-libav@jannau.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>
6 years agompegvideo: split ff_draw_horiz_band().
Anton Khirnov [Sun, 3 Feb 2013 13:26:34 +0000 (14:26 +0100)]
mpegvideo: split ff_draw_horiz_band().

Split out dependency on MpegEncContext.

6 years agoerror_resilience: decouple ER from MpegEncContext
Anton Khirnov [Sat, 2 Feb 2013 19:42:07 +0000 (20:42 +0100)]
error_resilience: decouple ER from MpegEncContext

6 years agosvq3: remove a pointless if()
Anton Khirnov [Mon, 4 Feb 2013 10:21:10 +0000 (11:21 +0100)]
svq3: remove a pointless if()

The H264 context is always uninitialized at this point.

6 years agoh264: remove a pointless if()
Anton Khirnov [Sun, 3 Feb 2013 09:21:39 +0000 (10:21 +0100)]
h264: remove a pointless if()

!encoding is always true, we do not have a H.264 encoder

6 years agoh264: simplify calls to ff_er_add_slice().
Anton Khirnov [Sat, 2 Feb 2013 14:18:00 +0000 (15:18 +0100)]
h264: simplify calls to ff_er_add_slice().

partitioned_frame is never set for h264 (as easily seen from git grep).

6 years agoget_buffer(): do not initialize the data.
Anton Khirnov [Wed, 30 Jan 2013 15:41:08 +0000 (16:41 +0100)]
get_buffer(): do not initialize the data.

There may be more decoders that rely on this. Those should be found and
fixed.

6 years agovf_yadif: fix out-of line reads
Anton Khirnov [Thu, 7 Feb 2013 17:31:45 +0000 (18:31 +0100)]
vf_yadif: fix out-of line reads

Some changes in the border pixels, visually indistinguishable.

6 years agovf_yadif: factorize initializing the filtering callbacks
Anton Khirnov [Sat, 9 Feb 2013 07:28:37 +0000 (08:28 +0100)]
vf_yadif: factorize initializing the filtering callbacks

Do it all in config_props().

6 years agosparc: dsputil: Simplify high_bit_depth checks
Diego Biurrun [Fri, 15 Feb 2013 01:13:42 +0000 (02:13 +0100)]
sparc: dsputil: Simplify high_bit_depth checks

6 years agoh264idct: Replace duplicate scan8 table by appropriate #include
Diego Biurrun [Fri, 18 Jan 2013 22:00:18 +0000 (23:00 +0100)]
h264idct: Replace duplicate scan8 table by appropriate #include

6 years agoavutil: Ensure that emms_c is always defined, even on non-x86
Diego Biurrun [Thu, 14 Feb 2013 18:07:21 +0000 (19:07 +0100)]
avutil: Ensure that emms_c is always defined, even on non-x86

6 years agoconfigure: Move MinGW CPPFLAGS setting to libc section, where it belongs
Diego Biurrun [Wed, 13 Feb 2013 19:00:47 +0000 (20:00 +0100)]
configure: Move MinGW CPPFLAGS setting to libc section, where it belongs

6 years agoavutil: Move emms code to x86-specific header
Diego Biurrun [Tue, 22 Jan 2013 01:22:29 +0000 (02:22 +0100)]
avutil: Move emms code to x86-specific header

6 years agodoc/platform: Fix 10l typo
Derek Buitenhuis [Wed, 13 Feb 2013 23:41:37 +0000 (18:41 -0500)]
doc/platform: Fix 10l typo

This error was somehow missed for months.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
6 years agodsputil: Move STRIDE_ALIGN macro to the only place it is used
Diego Biurrun [Fri, 8 Feb 2013 14:55:52 +0000 (15:55 +0100)]
dsputil: Move STRIDE_ALIGN macro to the only place it is used

6 years agolavr: fix mixing matrix reduction when normalization is disabled
Justin Ruggles [Sun, 27 Jan 2013 19:47:54 +0000 (14:47 -0500)]
lavr: fix mixing matrix reduction when normalization is disabled

In some cases when an input contributes fully to the corresponding
output, other inputs may also contribute to the same output. This is the
case, for example, for the default 5.1 to stereo downmix matrix without
normalization.

6 years agolavr: fix matrix reduction for upmixing in certain cases
Justin Ruggles [Fri, 18 Jan 2013 18:44:10 +0000 (13:44 -0500)]
lavr: fix matrix reduction for upmixing in certain cases

Do not skip an output if the corresponding input contributes to other output
channels.

6 years agolavr: cosmetics: reindent
Justin Ruggles [Fri, 18 Jan 2013 18:34:26 +0000 (13:34 -0500)]
lavr: cosmetics: reindent

6 years agolavr: make sure that the mix function is reset even if no mixing will be done
Justin Ruggles [Fri, 18 Jan 2013 18:27:50 +0000 (13:27 -0500)]
lavr: make sure that the mix function is reset even if no mixing will be done

If the matrix reduction ends up with no mixing matrix needed, we need to still
reset the mix function accordingly and log the info to the user.

6 years agolavr: print out the mix matrix in ff_audio_mix_set_matrix()
Justin Ruggles [Fri, 18 Jan 2013 18:25:19 +0000 (13:25 -0500)]
lavr: print out the mix matrix in ff_audio_mix_set_matrix()

This will print the new matrix if it is set after initialization.

6 years agows-snd1: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:31:24 +0000 (20:31 -0500)]
ws-snd1: decode directly to the user-provided AVFrame

6 years agowmavoice: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:29:44 +0000 (20:29 -0500)]
wmavoice: decode directly to the user-provided AVFrame

6 years agowmaenc: alloc/free coded_frame instead of keeping it in the WMACodecContext
Justin Ruggles [Mon, 24 Dec 2012 01:26:15 +0000 (20:26 -0500)]
wmaenc: alloc/free coded_frame instead of keeping it in the WMACodecContext

6 years agowma: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:25:52 +0000 (20:25 -0500)]
wma: decode directly to the user-provided AVFrame

6 years agowmapro: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:16:21 +0000 (20:16 -0500)]
wmapro: decode directly to the user-provided AVFrame

6 years agowavpack: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:13:08 +0000 (20:13 -0500)]
wavpack: decode directly to the user-provided AVFrame

6 years agovorbis: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:11:36 +0000 (20:11 -0500)]
vorbis: decode directly to the user-provided AVFrame

6 years agovmdaudio: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:10:01 +0000 (20:10 -0500)]
vmdaudio: decode directly to the user-provided AVFrame

6 years agotwinvq: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:08:05 +0000 (20:08 -0500)]
twinvq: decode directly to the user-provided AVFrame

6 years agotta: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:06:41 +0000 (20:06 -0500)]
tta: decode directly to the user-provided AVFrame

6 years agotruespeech: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:04:31 +0000 (20:04 -0500)]
truespeech: decode directly to the user-provided AVFrame

6 years agotak: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:02:46 +0000 (20:02 -0500)]
tak: decode directly to the user-provided AVFrame

6 years agosmackaud: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 01:00:41 +0000 (20:00 -0500)]
smackaud: decode directly to the user-provided AVFrame

6 years agosipr: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:58:35 +0000 (19:58 -0500)]
sipr: decode directly to the user-provided AVFrame

6 years agoshorten: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:56:32 +0000 (19:56 -0500)]
shorten: decode directly to the user-provided AVFrame

6 years agos302m: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:54:24 +0000 (19:54 -0500)]
s302m: decode directly to the user-provided AVFrame

6 years agora288: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:52:00 +0000 (19:52 -0500)]
ra288: decode directly to the user-provided AVFrame

6 years agora144: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:50:54 +0000 (19:50 -0500)]
ra144: decode directly to the user-provided AVFrame

6 years agoralf: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:49:38 +0000 (19:49 -0500)]
ralf: decode directly to the user-provided AVFrame

6 years agoqdm2: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:46:58 +0000 (19:46 -0500)]
qdm2: decode directly to the user-provided AVFrame

6 years agoqcelp: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:43:11 +0000 (19:43 -0500)]
qcelp: decode directly to the user-provided AVFrame

6 years agopcm-bluray: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:41:46 +0000 (19:41 -0500)]
pcm-bluray: decode directly to the user-provided AVFrame

6 years agonellymoser: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:39:44 +0000 (19:39 -0500)]
nellymoser: decode directly to the user-provided AVFrame

6 years agompc7/8: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:38:34 +0000 (19:38 -0500)]
mpc7/8: decode directly to the user-provided AVFrame

6 years agompegaudio: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:34:54 +0000 (19:34 -0500)]
mpegaudio: decode directly to the user-provided AVFrame

6 years agomlp/truehd: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:24:24 +0000 (19:24 -0500)]
mlp/truehd: decode directly to the user-provided AVFrame

6 years agomace: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:22:07 +0000 (19:22 -0500)]
mace: decode directly to the user-provided AVFrame

6 years agolibspeex: decode directly to the user-provided AVFrame
Justin Ruggles [Mon, 24 Dec 2012 00:19:25 +0000 (19:19 -0500)]
libspeex: decode directly to the user-provided AVFrame

6 years agolibopus: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:45:55 +0000 (18:45 -0500)]
libopus: decode directly to the user-provided AVFrame

6 years agolibopencore-amr: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:42:58 +0000 (18:42 -0500)]
libopencore-amr: decode directly to the user-provided AVFrame

6 years agolibgsm: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:40:50 +0000 (18:40 -0500)]
libgsm: decode directly to the user-provided AVFrame

6 years agolibilbc: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:05:53 +0000 (18:05 -0500)]
libilbc: decode directly to the user-provided AVFrame

6 years agodpcm: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:03:43 +0000 (18:03 -0500)]
dpcm: decode directly to the user-provided AVFrame

6 years agoimc/iac: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:01:45 +0000 (18:01 -0500)]
imc/iac: decode directly to the user-provided AVFrame

6 years agogsm: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 23:00:01 +0000 (18:00 -0500)]
gsm: decode directly to the user-provided AVFrame

6 years agog726: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:56:53 +0000 (17:56 -0500)]
g726: decode directly to the user-provided AVFrame

6 years agog723.1: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:55:09 +0000 (17:55 -0500)]
g723.1: decode directly to the user-provided AVFrame

6 years agog722: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:53:06 +0000 (17:53 -0500)]
g722: decode directly to the user-provided AVFrame

6 years agoflac: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:49:27 +0000 (17:49 -0500)]
flac: decode directly to the user-provided AVFrame

6 years agocinaudio: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:47:02 +0000 (17:47 -0500)]
cinaudio: decode directly to the user-provided AVFrame

6 years agodca: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:43:47 +0000 (17:43 -0500)]
dca: decode directly to the user-provided AVFrame

6 years agocook: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:41:31 +0000 (17:41 -0500)]
cook: decode directly to the user-provided AVFrame

6 years agocomfortnoise: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:37:34 +0000 (17:37 -0500)]
comfortnoise: decode directly to the user-provided AVFrame

6 years agobmvaudio: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:35:54 +0000 (17:35 -0500)]
bmvaudio: decode directly to the user-provided AVFrame

6 years agopcm: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:10:12 +0000 (17:10 -0500)]
pcm: decode directly to the user-provided AVFrame

6 years agobinkaudio: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:26:26 +0000 (17:26 -0500)]
binkaudio: decode directly to the user-provided AVFrame

6 years agoatrac3: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:24:53 +0000 (17:24 -0500)]
atrac3: decode directly to the user-provided AVFrame

6 years agoatrac1: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:07:10 +0000 (17:07 -0500)]
atrac1: decode directly to the user-provided AVFrame

6 years agoape: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:05:40 +0000 (17:05 -0500)]
ape: decode directly to the user-provided AVFrame

6 years agoamrwb: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:03:36 +0000 (17:03 -0500)]
amrwb: decode directly to the user-provided AVFrame

6 years agoamrnb: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 22:01:33 +0000 (17:01 -0500)]
amrnb: decode directly to the user-provided AVFrame

6 years agoals: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:59:43 +0000 (16:59 -0500)]
als: decode directly to the user-provided AVFrame

6 years agoalac: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:56:17 +0000 (16:56 -0500)]
alac: decode directly to the user-provided AVFrame

6 years agoadxenc: alloc/free coded_frame instead of keeping it in the ADXContext
Justin Ruggles [Sun, 23 Dec 2012 21:51:41 +0000 (16:51 -0500)]
adxenc: alloc/free coded_frame instead of keeping it in the ADXContext

6 years agoadx: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:50:06 +0000 (16:50 -0500)]
adx: decode directly to the user-provided AVFrame

6 years agoadpcm: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:43:07 +0000 (16:43 -0500)]
adpcm: decode directly to the user-provided AVFrame

6 years agoac3: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:40:37 +0000 (16:40 -0500)]
ac3: decode directly to the user-provided AVFrame

6 years agoaac: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:37:23 +0000 (16:37 -0500)]
aac: decode directly to the user-provided AVFrame

6 years ago8svx: decode directly to the user-provided AVFrame
Justin Ruggles [Sun, 23 Dec 2012 21:28:03 +0000 (16:28 -0500)]
8svx: decode directly to the user-provided AVFrame

6 years agolavu: avoid clashing definition of E
Luca Barbato [Sat, 9 Feb 2013 21:37:07 +0000 (22:37 +0100)]
lavu: avoid clashing definition of E

E is usually defined as a shorthand for AV_OPT_FLAG_ENCODING_PARAM.
Rename the single expansion E(x) now used in libavutil to E1.

6 years agodoc: developer: Add a note about reserved system name space
Diego Biurrun [Mon, 11 Feb 2013 17:53:33 +0000 (18:53 +0100)]
doc: developer: Add a note about reserved system name space

6 years agodoc: developer: Fix wording in "naming conventions" section
Diego Biurrun [Mon, 11 Feb 2013 17:39:12 +0000 (18:39 +0100)]
doc: developer: Fix wording in "naming conventions" section

6 years agoopenbsd: Add minor number to shared library install name
Diego Biurrun [Sun, 3 Feb 2013 14:02:31 +0000 (15:02 +0100)]
openbsd: Add minor number to shared library install name

This is what the OpenBSD porter's manual describes as correct, cf.
http://www.openbsd.org/faq/ports/specialtopics.html#SharedLibs

6 years agobuild: Make library minor version visible in the Makefile
Diego Biurrun [Sun, 3 Feb 2013 14:03:08 +0000 (15:03 +0100)]
build: Make library minor version visible in the Makefile

This allows employing that number in library install commands.

6 years agox86: mpeg4qpel: Make movsxifnidn do the right thing
Daniel Kang [Mon, 11 Feb 2013 18:45:10 +0000 (13:45 -0500)]
x86: mpeg4qpel: Make movsxifnidn do the right thing

Fixes an instruction that does nothing by changing the
source to dword.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
6 years agoh264: Copy h264chroma dsp context to slice thread copies
Martin Storsjö [Mon, 11 Feb 2013 18:27:50 +0000 (20:27 +0200)]
h264: Copy h264chroma dsp context to slice thread copies

This fixes slice threading which seems to have been broken since
79dad2a93.

Signed-off-by: Martin Storsjö <martin@martin.st>
6 years agoswscale: Disallow conversion to GBRP16
Martin Storsjö [Mon, 11 Feb 2013 11:37:29 +0000 (13:37 +0200)]
swscale: Disallow conversion to GBRP16

This reverts parts of d6d5ef5534d582, that didn't work right. (The
tests that were added failed on big endian, and the output looked
garbled on little endian as well.)

This is due to the fact that the intermediate scaling values (from
e.g. hScale8To19_c or hScale16To19_c) are stored as int32_t and
thus requires a separate output function, while yuv2gbrp_full_X_c
only interprets it as int16_t.

Signed-off-by: Martin Storsjö <martin@martin.st>
6 years agocrc: Move static data table from header to .c file
Diego Biurrun [Sun, 3 Feb 2013 14:43:47 +0000 (15:43 +0100)]
crc: Move static data table from header to .c file

Having static data tables in header files is a potential source of trouble.

6 years agoavcodec/rectangle: Remove nonsense assert
Diego Biurrun [Fri, 8 Feb 2013 14:55:24 +0000 (15:55 +0100)]
avcodec/rectangle: Remove nonsense assert

6 years agolibfdk-aacenc: Actually check for upper bounds of cutoff
Derek Buitenhuis [Sat, 9 Feb 2013 19:03:02 +0000 (14:03 -0500)]
libfdk-aacenc: Actually check for upper bounds of cutoff

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
6 years agoarm: Fall back to runtime cpu feature detection via /proc/cpuinfo
Martin Storsjö [Thu, 7 Feb 2013 08:54:20 +0000 (10:54 +0200)]
arm: Fall back to runtime cpu feature detection via /proc/cpuinfo

On recent android versions, /proc/self/auxw is unreadable
(unless the process is running running under the shell uid or
in debuggable mode, which makes it hard to notice). See
http://b.android.com/43055 and
https://android-review.googlesource.com/51271 for more information
about the issue.

This makes sure e.g. neon optimizations are enabled at runtime in
android apps even when built in release mode, if configured to
use the runtime detection.

CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>