platform/upstream/libav.git
14 years agoindent
Michael Niedermayer [Fri, 24 Sep 2010 16:18:40 +0000 (16:18 +0000)]
indent

Originally committed as revision 25178 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetic (rename detect_repeats to is_atty which matches the truth)
Michael Niedermayer [Fri, 24 Sep 2010 16:18:10 +0000 (16:18 +0000)]
Cosmetic (rename detect_repeats to is_atty which matches the truth)

Originally committed as revision 25177 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoDoxy consistency cosmetics
Michael Niedermayer [Fri, 24 Sep 2010 15:40:53 +0000 (15:40 +0000)]
Doxy consistency cosmetics

Originally committed as revision 25176 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoEnable AV_LOG_SKIP_REPEATED to maintain previous behavior.
Michael Niedermayer [Fri, 24 Sep 2010 15:39:10 +0000 (15:39 +0000)]
Enable AV_LOG_SKIP_REPEATED to maintain previous behavior.

Originally committed as revision 25175 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years ago2nd try to fix av_log() repeated detection
Michael Niedermayer [Fri, 24 Sep 2010 15:37:01 +0000 (15:37 +0000)]
2nd try to fix av_log() repeated detection

Originally committed as revision 25174 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove unused variable.
Ronald S. Bultje [Fri, 24 Sep 2010 15:31:46 +0000 (15:31 +0000)]
Remove unused variable.

Originally committed as revision 25173 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUnroll loop in h264_idct_add16intra_sse2(). Basically identical to r25171, this
Ronald S. Bultje [Fri, 24 Sep 2010 14:07:23 +0000 (14:07 +0000)]
Unroll loop in h264_idct_add16intra_sse2(). Basically identical to r25171, this
inlines scan8[] and removes loop setup. 15% faster, 0.4% overall.

See "[PATCH] unroll loop in h264_idct_add8_sse2()" thread on ML.

Originally committed as revision 25172 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUnroll loop in h264_idct_add8_sse2(). This means we can inline scan8[] in the
Ronald S. Bultje [Fri, 24 Sep 2010 14:05:45 +0000 (14:05 +0000)]
Unroll loop in h264_idct_add8_sse2(). This means we can inline scan8[] in the
code directly also and remove loop setup. 20% faster in function, 0.8% overall.

See "[PATCH] unroll loop in h264_idct_add8_sse2()" thread on ML.

Originally committed as revision 25171 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUpdate docs regarding writing optimizations:
Ronald S. Bultje [Fri, 24 Sep 2010 14:01:09 +0000 (14:01 +0000)]
Update docs regarding writing optimizations:
- mention clobber-marking of xmm registers,
- some notes on external vs. inline asm, including tips on which to use for
   what situation and to not rewrite+improve in the same patch (as with C code)
- some more best-practice guidelines

See "[PATCH] update doc/optimization.txt" thread on ML.

Originally committed as revision 25170 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove unnecessary av_strdup() and av_free().
Stefano Sabatini [Fri, 24 Sep 2010 09:15:27 +0000 (09:15 +0000)]
Remove unnecessary av_strdup() and av_free().

Originally committed as revision 25169 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUpdate doc/APIchanges after the last API changes.
Stefano Sabatini [Fri, 24 Sep 2010 00:55:02 +0000 (00:55 +0000)]
Update doc/APIchanges after the last API changes.

Originally committed as revision 25168 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMake ffplay -h show the settable AVOptions.
Stefano Sabatini [Fri, 24 Sep 2010 00:51:43 +0000 (00:51 +0000)]
Make ffplay -h show the settable AVOptions.

Originally committed as revision 25167 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoDeprecate av_opt_show() in favor of a new function av_opt_show2(),
Stefano Sabatini [Fri, 24 Sep 2010 00:51:40 +0000 (00:51 +0000)]
Deprecate av_opt_show() in favor of a new function av_opt_show2(),
which allows to specify only a subset of all the options to show.

Originally committed as revision 25166 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd frei0r filter.
Stefano Sabatini [Fri, 24 Sep 2010 00:32:22 +0000 (00:32 +0000)]
Add frei0r filter.

See thread:
Subject: [FFmpeg-devel] [POC] frei0r wrapper
Date: Tue, 24 Aug 2010 21:37:32 +0200

Originally committed as revision 25165 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd a G.722 encoder
Martin Storsjö [Thu, 23 Sep 2010 21:33:29 +0000 (21:33 +0000)]
Add a G.722 encoder

Originally committed as revision 25164 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agorawdec: only allocate a full-frame size buffer if it actually will
Reimar Döffinger [Thu, 23 Sep 2010 20:23:15 +0000 (20:23 +0000)]
rawdec: only allocate a full-frame size buffer if it actually will
be used, place palette buffer in the context to simplify this.

Originally committed as revision 25163 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCorrect terminology bug in poll_frame()
Michael Niedermayer [Thu, 23 Sep 2010 11:12:03 +0000 (11:12 +0000)]
Correct terminology bug in poll_frame()
it returns the number of samples not frames (for video sample=frame)

Originally committed as revision 25162 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agofate: print commands being executed with V=1
Måns Rullgård [Thu, 23 Sep 2010 09:55:31 +0000 (09:55 +0000)]
fate: print commands being executed with V=1

Originally committed as revision 25161 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix indentation after r25158.
Carl Eugen Hoyos [Thu, 23 Sep 2010 09:53:26 +0000 (09:53 +0000)]
Fix indentation after r25158.

Originally committed as revision 25160 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAllow to set the frame rate in v4l2 devices
José Miguel Gonçalves [Thu, 23 Sep 2010 09:16:05 +0000 (09:16 +0000)]
Allow to set the frame rate in v4l2 devices
Patch by José Miguel Gonçalves (jose DOT goncalves AT inov DOT pt)

Originally committed as revision 25159 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix aspect ratio for files that have it stored in
Richard Buteau [Thu, 23 Sep 2010 03:40:06 +0000 (03:40 +0000)]
Fix aspect ratio for files that have it stored in
ff_asf_extended_content_header.
Fixes issue 690.

Patch by Richard Buteau, rbuteau rgbnetworks com

Originally committed as revision 25158 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix typo (instruciton -> instruction).
Ronald S. Bultje [Wed, 22 Sep 2010 02:18:03 +0000 (02:18 +0000)]
Fix typo (instruciton -> instruction).

Originally committed as revision 25157 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoThe debug text output of macroblocks can indicate MB_TYPE_INTERLACED,
Ivan Kalvachev [Tue, 21 Sep 2010 22:44:27 +0000 (22:44 +0000)]
The debug text output of macroblocks can indicate MB_TYPE_INTERLACED,
but it used to do it only for h264 codec.
Allow it for other codecs, as mpeg2 and mpeg4 also set this flag.

Originally committed as revision 25156 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agocosmetics: Place concat protocol entry in alphabetical order.
Diego Biurrun [Tue, 21 Sep 2010 22:21:12 +0000 (22:21 +0000)]
cosmetics: Place concat protocol entry in alphabetical order.

Originally committed as revision 25155 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agotcp: Check both wfds and efds when waiting for the result from connect
Martin Storsjö [Tue, 21 Sep 2010 20:17:34 +0000 (20:17 +0000)]
tcp: Check both wfds and efds when waiting for the result from connect

On windows, a connection failure doesn't trigger wfds as it does on unix.

This fixes issue 2237, based on code by yeyingxian.

Originally committed as revision 25154 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agox86: disable SSE functions using stack when stack is not aligned
Måns Rullgård [Tue, 21 Sep 2010 17:57:21 +0000 (17:57 +0000)]
x86: disable SSE functions using stack when stack is not aligned

This fixes crashes with ICC 10.1.

Originally committed as revision 25153 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoDisable ALIGNED_STACK with icc 10 or prior on x86_32
Måns Rullgård [Tue, 21 Sep 2010 17:57:17 +0000 (17:57 +0000)]
Disable ALIGNED_STACK with icc 10 or prior on x86_32

Originally committed as revision 25152 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd HAVE_ALIGNED_STACK config setting
Måns Rullgård [Tue, 21 Sep 2010 17:57:13 +0000 (17:57 +0000)]
Add HAVE_ALIGNED_STACK config setting

This is set to 1 if the stack is guaranteed to be suitably aligned
for the strictest access mode of the machine.

Originally committed as revision 25151 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoARM: disable movw/movt for relocated values on Apple platforms
Måns Rullgård [Mon, 20 Sep 2010 21:50:46 +0000 (21:50 +0000)]
ARM: disable movw/movt for relocated values on Apple platforms

Apparently Apple platforms do not handle movw/movt relocations
properly, leading to runtime crashes in code using them.

Originally committed as revision 25150 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove log_callback_help to cmdutils.[hc], for allowing sharing.
Stefano Sabatini [Mon, 20 Sep 2010 21:48:03 +0000 (21:48 +0000)]
Move log_callback_help to cmdutils.[hc], for allowing sharing.

Originally committed as revision 25149 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoSupport deinterlacing of YUVJ420P.
Benjamin Larsson [Mon, 20 Sep 2010 11:34:27 +0000 (11:34 +0000)]
Support deinterlacing of YUVJ420P.

Originally committed as revision 25148 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCheck for the IPPROTO_IPV6 define before using it
Martin Storsjö [Mon, 20 Sep 2010 06:38:40 +0000 (06:38 +0000)]
Check for the IPPROTO_IPV6 define before using it

This fixes building on FreeBSD in some configurations, if the IPv6 multicast
structs are available, but IPPROTO_IPV6 isn't defined.

Originally committed as revision 25147 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agox86: remove hack disabling sse2 h264 loop filter with 32-bit icc
Måns Rullgård [Sat, 18 Sep 2010 20:44:32 +0000 (20:44 +0000)]
x86: remove hack disabling sse2 h264 loop filter with 32-bit icc

Originally committed as revision 25146 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRequest 16-byte aligned stack with icc on x86_32
Måns Rullgård [Sat, 18 Sep 2010 20:44:29 +0000 (20:44 +0000)]
Request 16-byte aligned stack with icc on x86_32

Originally committed as revision 25145 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years ago10l: error in last commit. use decoder channels not encoder channels.
Justin Ruggles [Sat, 18 Sep 2010 20:19:18 +0000 (20:19 +0000)]
10l: error in last commit. use decoder channels not encoder channels.

Originally committed as revision 25144 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoChange remaining ost->st->codec and ist->st->codec to enc and dec in
Justin Ruggles [Sat, 18 Sep 2010 20:16:56 +0000 (20:16 +0000)]
Change remaining ost->st->codec and ist->st->codec to enc and dec in
do_audio_out().

Originally committed as revision 25143 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove legacy x264 presets
Lou Logan [Fri, 17 Sep 2010 20:14:24 +0000 (20:14 +0000)]
Remove legacy x264 presets

Since we now have the official x264 presets in ffmpeg, there's no reason to
keep around the old ones.

Patch by Lou Logan <lou AT fakeoutdoorsman DOT com>.

Originally committed as revision 25142 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove pointless semicolon
Vitor Sessak [Fri, 17 Sep 2010 19:33:56 +0000 (19:33 +0000)]
Remove pointless semicolon

Originally committed as revision 25141 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoDon't access upper 32 bits of a 32-bit int on 64-bit systems.
Ronald S. Bultje [Fri, 17 Sep 2010 12:24:22 +0000 (12:24 +0000)]
Don't access upper 32 bits of a 32-bit int on 64-bit systems.

Originally committed as revision 25140 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoconfigure: print minimum lame version number required after revision 25128
James Darnley [Fri, 17 Sep 2010 10:52:43 +0000 (10:52 +0000)]
configure: print minimum lame version number required after revision 25128

Patch by James Darnley <james dot darnley at gmail dot com>.

Originally committed as revision 25139 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoProperly add HAVE_YASM around yasmified symbols. Should fix compile error
Ronald S. Bultje [Fri, 17 Sep 2010 03:01:57 +0000 (03:01 +0000)]
Properly add HAVE_YASM around yasmified symbols. Should fix compile error
on configurations using --disable-yasm.

Originally committed as revision 25138 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove hadamard_diff{,16}_{mmx,mmx2,sse2,ssse3}() from inline asm to yasm,
Ronald S. Bultje [Fri, 17 Sep 2010 01:56:06 +0000 (01:56 +0000)]
Move hadamard_diff{,16}_{mmx,mmx2,sse2,ssse3}() from inline asm to yasm,
which will hopefully solve the Win64/FATE failures caused by these functions.

Originally committed as revision 25137 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove sse16_sse2() from inline asm to yasm. It is one of the functions causing
Ronald S. Bultje [Fri, 17 Sep 2010 01:44:17 +0000 (01:44 +0000)]
Move sse16_sse2() from inline asm to yasm. It is one of the functions causing
Win64/FATE issues.

Originally committed as revision 25136 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoconfigure: detect Open64 compiler
Måns Rullgård [Thu, 16 Sep 2010 15:25:10 +0000 (15:25 +0000)]
configure: detect Open64 compiler

Originally committed as revision 25135 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoswscale-test: cosmetic alignment
Ramiro Polla [Thu, 16 Sep 2010 13:30:02 +0000 (13:30 +0000)]
swscale-test: cosmetic alignment

Originally committed as revision 32265 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoCosmetics: fix weird align.
Stefano Sabatini [Thu, 16 Sep 2010 08:47:37 +0000 (08:47 +0000)]
Cosmetics: fix weird align.

Originally committed as revision 25134 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUse <> for system headers inclusion.
Stefano Sabatini [Thu, 16 Sep 2010 08:47:34 +0000 (08:47 +0000)]
Use <> for system headers inclusion.

Originally committed as revision 25133 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix copyright notice, make it more consistent with the rest of FFmpeg.
Stefano Sabatini [Thu, 16 Sep 2010 08:47:31 +0000 (08:47 +0000)]
Fix copyright notice, make it more consistent with the rest of FFmpeg.

Originally committed as revision 25132 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix grammar in the ocv_smooth filter documentation.
Stefano Sabatini [Thu, 16 Sep 2010 08:37:15 +0000 (08:37 +0000)]
Fix grammar in the ocv_smooth filter documentation.

Originally committed as revision 25131 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoPrefer "X" over ``X'', looks more readable and more consistent with
Stefano Sabatini [Thu, 16 Sep 2010 08:34:55 +0000 (08:34 +0000)]
Prefer "X" over ``X'', looks more readable and more consistent with
the rest of FFmpeg docs.

Originally committed as revision 25130 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRemove useless comment.
Carl Eugen Hoyos [Thu, 16 Sep 2010 08:30:01 +0000 (08:30 +0000)]
Remove useless comment.

Originally committed as revision 25129 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoswscale: remove useless temporary variable
Ramiro Polla [Thu, 16 Sep 2010 01:09:59 +0000 (01:09 +0000)]
swscale: remove useless temporary variable

Originally committed as revision 32258 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale: factorize plane copying code out of 2 functions
Ramiro Polla [Thu, 16 Sep 2010 01:07:55 +0000 (01:07 +0000)]
swscale: factorize plane copying code out of 2 functions

Originally committed as revision 32257 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale-test: always use bilinear scaler to get output for SSD
Ramiro Polla [Thu, 16 Sep 2010 00:41:08 +0000 (00:41 +0000)]
swscale-test: always use bilinear scaler to get output for SSD

Originally committed as revision 32256 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoAllow float values for libmp3lame quality.
James Darnley [Wed, 15 Sep 2010 22:10:13 +0000 (22:10 +0000)]
Allow float values for libmp3lame quality.

Patch by James Darnley, james D darnley A gmail

Originally committed as revision 25128 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAllow audiogen to take commandline parameters for sample rate and number of
Justin Ruggles [Wed, 15 Sep 2010 21:36:04 +0000 (21:36 +0000)]
Allow audiogen to take commandline parameters for sample rate and number of
channels.

Originally committed as revision 25127 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agortsp: Handle standard assigned codec names for private payload types, too
Martin Storsjö [Wed, 15 Sep 2010 17:39:25 +0000 (17:39 +0000)]
rtsp: Handle standard assigned codec names for private payload types, too

Originally committed as revision 25126 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoHandle G.722 in RTP, and all the exceptions mandated in RFC 3551
Martin Storsjö [Wed, 15 Sep 2010 17:35:39 +0000 (17:35 +0000)]
Handle G.722 in RTP, and all the exceptions mandated in RFC 3551

Originally committed as revision 25125 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCheck avctx width/height more thoroughly (e.g. all values 0 except width would
Reimar Döffinger [Wed, 15 Sep 2010 04:46:55 +0000 (04:46 +0000)]
Check avctx width/height more thoroughly (e.g. all values 0 except width would
have been accepted before).
Also do not fail if they are invalid but instead override them to 0.
This allows decoding e.g. MPEG video when only the container values are corrupted.
For encoding a value of 0,0 of course makes no sense, but was allowed
through before and will be caught by an extra check in the encode function.

Originally committed as revision 25124 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoClarify what av_read_frame() returns.
Michael Niedermayer [Tue, 14 Sep 2010 22:20:46 +0000 (22:20 +0000)]
Clarify what av_read_frame() returns.

Originally committed as revision 25123 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUpdate pathscale compiler detection
Måns Rullgård [Tue, 14 Sep 2010 18:55:20 +0000 (18:55 +0000)]
Update pathscale compiler detection

Originally committed as revision 25122 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoARM: fix NEON h264_idct_add8
Måns Rullgård [Tue, 14 Sep 2010 17:11:51 +0000 (17:11 +0000)]
ARM: fix NEON h264_idct_add8

Originally committed as revision 25121 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd av_popcount() to libavutil/common.h and bump minor version
Tomas Härdin [Tue, 14 Sep 2010 14:45:43 +0000 (14:45 +0000)]
Add av_popcount() to libavutil/common.h and bump minor version

Originally committed as revision 25120 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRename h264_idct_sse2.asm to h264_idct.asm; move inline IDCT asm from
Ronald S. Bultje [Tue, 14 Sep 2010 13:36:26 +0000 (13:36 +0000)]
Rename h264_idct_sse2.asm to h264_idct.asm; move inline IDCT asm from
h264dsp_mmx.c to h264_idct.asm (as yasm code). Because the loops are now
coded in asm instead of C, this is (depending on the function) up to 50%
faster for cases where gcc didn't do a great job at looping.

Since h264_idct_add8() is now faster than the manual loop setup in h264.c,
in-asm idct calling can now be enabled for chroma as well (see r16207). For
MMX, this is 5% faster. For SSE2 (which isn't done for chroma if h264.c does
the looping), this makes it up to 50% faster. Speed gain overall is ~0.5-1.0%.

Originally committed as revision 25119 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoImplement libopencv smooth filter.
Stefano Sabatini [Tue, 14 Sep 2010 13:21:13 +0000 (13:21 +0000)]
Implement libopencv smooth filter.

Originally committed as revision 25118 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agorgb2rgb: don't misuse HAVE_* defines
Ramiro Polla [Tue, 14 Sep 2010 13:12:11 +0000 (13:12 +0000)]
rgb2rgb: don't misuse HAVE_* defines

Introduce and use COMPILE_TEMPLATE_* instead.

Originally committed as revision 32241 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoLimit av_log repeat detection to terminals so as to avoid filling files with
Michael Niedermayer [Tue, 14 Sep 2010 00:17:58 +0000 (00:17 +0000)]
Limit av_log repeat detection to terminals so as to avoid filling files with
lots of mess.

Originally committed as revision 25117 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoReindent after r25115.
Carl Eugen Hoyos [Mon, 13 Sep 2010 22:09:28 +0000 (22:09 +0000)]
Reindent after r25115.

Originally committed as revision 25116 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAdd R10k decoder.
Zhou Zongyi [Mon, 13 Sep 2010 22:08:51 +0000 (22:08 +0000)]
Add R10k decoder.

Original patch by Zhou Zongyi, zhouzy A os pku edu cn, resubmitted by
James Darnley, james.darnley gmail, changes by me.

Originally committed as revision 25115 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoTest lowres before codec init.
Carl Eugen Hoyos [Mon, 13 Sep 2010 22:03:29 +0000 (22:03 +0000)]
Test lowres before codec init.

Originally committed as revision 25114 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoswscale: indentation and emtpy line cosmetics
Ramiro Polla [Mon, 13 Sep 2010 14:33:08 +0000 (14:33 +0000)]
swscale: indentation and emtpy line cosmetics

Originally committed as revision 32222 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale: avoid reading prior to the source buffer in planar2x() MMX2
Ramiro Polla [Mon, 13 Sep 2010 14:25:18 +0000 (14:25 +0000)]
swscale: avoid reading prior to the source buffer in planar2x() MMX2

Originally committed as revision 32221 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agofix x11grab example in e.g. the manpage so that they actually work
Reinhard Tartler [Mon, 13 Sep 2010 13:25:51 +0000 (13:25 +0000)]
fix x11grab example in e.g. the manpage so that they actually work

Originally committed as revision 25113 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoApply minor simplification in av_read_image_line().
Stefano Sabatini [Sun, 12 Sep 2010 22:00:45 +0000 (22:00 +0000)]
Apply minor simplification in av_read_image_line().

Originally committed as revision 25112 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agopixdesc: indent
Måns Rullgård [Sun, 12 Sep 2010 21:39:57 +0000 (21:39 +0000)]
pixdesc: indent

Originally committed as revision 25111 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agopixdesc: use 8-bit accesses when possible in av_read/write_image_line()
Måns Rullgård [Sun, 12 Sep 2010 21:39:54 +0000 (21:39 +0000)]
pixdesc: use 8-bit accesses when possible in av_read/write_image_line()

This fixes out of bounds accesses for big endian formats and should be
a little faster.

Originally committed as revision 25110 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetics: use x ? y : z construct and merge intialization and
Stefano Sabatini [Sun, 12 Sep 2010 21:31:39 +0000 (21:31 +0000)]
Cosmetics: use x ? y : z construct and merge intialization and
definition for the val variable in av_read_image_line(). Simplify.

Originally committed as revision 25109 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetics: simplify layout of fmt_conversion_table, make it shorter
Stefano Sabatini [Sun, 12 Sep 2010 19:44:45 +0000 (19:44 +0000)]
Cosmetics: simplify layout of fmt_conversion_table, make it shorter
and more readable.

Originally committed as revision 25108 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoswscale-test: allow specifying src/dst pixel formats
Ramiro Polla [Sun, 12 Sep 2010 18:17:29 +0000 (18:17 +0000)]
swscale-test: allow specifying src/dst pixel formats

Originally committed as revision 32215 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoindent
Ramiro Polla [Sun, 12 Sep 2010 18:16:53 +0000 (18:16 +0000)]
indent

Originally committed as revision 32214 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale-test: take input from reference file
Ramiro Polla [Sun, 12 Sep 2010 18:16:33 +0000 (18:16 +0000)]
swscale-test: take input from reference file

If the CRC from the src->dst conversion matches a reference, it is not
necessary to perform a dst->yuva420p conversion and check the SSD.

Originally committed as revision 32213 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale-test: always use SWS_BILINEAR to convert from ref to src
Ramiro Polla [Sun, 12 Sep 2010 18:15:34 +0000 (18:15 +0000)]
swscale-test: always use SWS_BILINEAR to convert from ref to src

Originally committed as revision 32212 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale-test: move conversion from ref to source back to doTest()
Ramiro Polla [Sun, 12 Sep 2010 18:14:42 +0000 (18:14 +0000)]
swscale-test: move conversion from ref to source back to doTest()

The source format parameters are kept in static variables and conversion from
ref to source is only made when any parameter changes.

Originally committed as revision 32211 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agorgb2rgb: replace shuffle_bytes_2103() by optimized rgb32tobgr32()
Ramiro Polla [Sun, 12 Sep 2010 13:03:41 +0000 (13:03 +0000)]
rgb2rgb: replace shuffle_bytes_2103() by optimized rgb32tobgr32()

Originally committed as revision 32190 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoSet a constant frame size for encoding G.726 audio.
Justin Ruggles [Sat, 11 Sep 2010 19:52:09 +0000 (19:52 +0000)]
Set a constant frame size for encoding G.726 audio.

Originally committed as revision 25107 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetics: apply minor style fixes.
Stefano Sabatini [Sat, 11 Sep 2010 16:44:10 +0000 (16:44 +0000)]
Cosmetics: apply minor style fixes.

Originally committed as revision 25106 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoRead all id3v2 tags at the beginning of mp3 files.
David Byron [Sat, 11 Sep 2010 16:29:23 +0000 (16:29 +0000)]
Read all id3v2 tags at the beginning of mp3 files.

Patch by David Byron, dbyron dbyron com

Originally committed as revision 25105 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoindent
Ramiro Polla [Sat, 11 Sep 2010 14:31:13 +0000 (14:31 +0000)]
indent

Originally committed as revision 32157 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoswscale: remove unused code
Ramiro Polla [Sat, 11 Sep 2010 14:30:54 +0000 (14:30 +0000)]
swscale: remove unused code

yvu9ToYv12Wrapper() used to support yv12 with the chroma planes either in the
uv order or the vu order. FFmpeg no longer has a pixel format in vu order.

Originally committed as revision 32156 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agorgb2rgb: remove unused yvu9toyv12 function
Ramiro Polla [Sat, 11 Sep 2010 14:29:55 +0000 (14:29 +0000)]
rgb2rgb: remove unused yvu9toyv12 function

Originally committed as revision 32155 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

14 years agoAdd a FATE test for the G.722 decoder
Martin Storsjö [Sat, 11 Sep 2010 10:23:02 +0000 (10:23 +0000)]
Add a FATE test for the G.722 decoder

Originally committed as revision 25104 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agotta: remove stray semicolon
Måns Rullgård [Sat, 11 Sep 2010 09:32:14 +0000 (09:32 +0000)]
tta: remove stray semicolon

Originally committed as revision 25103 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoAllow mpeg encoding with qscale and very low frame rate.
James Darnley [Fri, 10 Sep 2010 23:41:14 +0000 (23:41 +0000)]
Allow mpeg encoding with qscale and very low frame rate.

Patch by James Darnley, james D darnley A gmail

Originally committed as revision 25102 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFLV Metadata
Tomás Touceda [Fri, 10 Sep 2010 23:29:07 +0000 (23:29 +0000)]
FLV Metadata

Patch by Tomás Touceda, chiiph gentoo org

Originally committed as revision 25101 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoMove AMR-NB frame unpacking code to a common file so it can be reused in
Marcelo Galvăo Póvoa [Fri, 10 Sep 2010 19:51:08 +0000 (19:51 +0000)]
Move AMR-NB frame unpacking code to a common file so it can be reused in
the AMR-WB decoder.

Patch by Marcelo Galvăo Póvoa.

Originally committed as revision 25100 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoUpdate H263_AIC asm offset for the apple variant
Luca Barbato [Fri, 10 Sep 2010 19:25:42 +0000 (19:25 +0000)]
Update H263_AIC asm offset for the apple variant

Originally committed as revision 25099 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoaacdec: Rework channel mapping compatibility hacks.
Alex Converse [Fri, 10 Sep 2010 18:01:48 +0000 (18:01 +0000)]
aacdec: Rework channel mapping compatibility hacks.

For a PCE based configuration map the channels solely based on tags.
For an indexed configuration map the channels solely based on position.

This works with all known exotic samples including al17, elem_id0, bad_concat,
and lfe_is_sce.

Originally committed as revision 25098 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoCosmetics: remove useless parentheses.
Stefano Sabatini [Fri, 10 Sep 2010 17:44:08 +0000 (17:44 +0000)]
Cosmetics: remove useless parentheses.

Originally committed as revision 25097 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agoFix 24 bpp CSCD decoding, as for Windows bitmaps in this (and only this)
Reimar Döffinger [Fri, 10 Sep 2010 17:33:31 +0000 (17:33 +0000)]
Fix 24 bpp CSCD decoding, as for Windows bitmaps in this (and only this)
case the stride must be aligned to a multiple of 4.
The original CSCD encoder just compresses bitmaps it gets via Windows
API functions as-is, thus it uses exactly those alignment rules.

Originally committed as revision 25096 to svn://svn.ffmpeg.org/ffmpeg/trunk

14 years agonutenc: fix unstable floating-point calculations
Måns Rullgård [Fri, 10 Sep 2010 13:16:08 +0000 (13:16 +0000)]
nutenc: fix unstable floating-point calculations

Originally committed as revision 25095 to svn://svn.ffmpeg.org/ffmpeg/trunk