platform/upstream/libav.git
16 years agomakes really sure that no undefined behavior can happen
Aurelien Jacobs [Wed, 27 Feb 2008 00:53:34 +0000 (00:53 +0000)]
makes really sure that no undefined behavior can happen

Originally committed as revision 12258 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agobetter include
Michael Niedermayer [Tue, 26 Feb 2008 23:51:59 +0000 (23:51 +0000)]
better include

Originally committed as revision 12257 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoinclude dependencies ...
Michael Niedermayer [Tue, 26 Feb 2008 23:49:41 +0000 (23:49 +0000)]
include dependencies ...

Originally committed as revision 12256 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAdd multiple inclusion guards.
Diego Biurrun [Tue, 26 Feb 2008 23:17:08 +0000 (23:17 +0000)]
Add multiple inclusion guards.

Originally committed as revision 12255 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAdd explanatory comment to '#define FLOAT float'.
Diego Biurrun [Tue, 26 Feb 2008 22:52:01 +0000 (22:52 +0000)]
Add explanatory comment to '#define FLOAT float'.

Originally committed as revision 12254 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAvoid infinite loop.
Aurelien Jacobs [Tue, 26 Feb 2008 22:35:07 +0000 (22:35 +0000)]
Avoid infinite loop.
uint64_t >> 64 is an undefined operation

Originally committed as revision 12253 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoMake gcc ignore the distinction between +0.0 and -0.0, this allows it to
Michael Niedermayer [Tue, 26 Feb 2008 22:01:11 +0000 (22:01 +0000)]
Make gcc ignore the distinction between +0.0 and -0.0, this allows it to
change x+0.0 to x. Though it alone does not allow gcc to remove x*0.0,
that would also require -ffinite-math-only.

Originally committed as revision 12252 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoprint enabled things in columns
Måns Rullgård [Tue, 26 Feb 2008 21:50:06 +0000 (21:50 +0000)]
print enabled things in columns

Originally committed as revision 12251 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago10l, alphabetic ordering
Vitor Sessak [Tue, 26 Feb 2008 21:34:13 +0000 (21:34 +0000)]
10l, alphabetic ordering

Originally committed as revision 12250 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoBuild system changes for libavfilter addition. Add --enable-avfilter configure option.
Vitor Sessak [Tue, 26 Feb 2008 21:32:56 +0000 (21:32 +0000)]
Build system changes for libavfilter addition. Add --enable-avfilter configure option.

Patch by myself, Bobby Bingham and Víctor Paesa.

Originally committed as revision 12249 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoMake libavfilter define its version like the other libav* libs
Vitor Sessak [Tue, 26 Feb 2008 21:21:17 +0000 (21:21 +0000)]
Make libavfilter define its version like the other libav* libs

Originally committed as revision 12248 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosimplify version number extraction
Måns Rullgård [Tue, 26 Feb 2008 21:18:55 +0000 (21:18 +0000)]
simplify version number extraction

Originally committed as revision 12247 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoless preprocessor magic in version number macros
Måns Rullgård [Tue, 26 Feb 2008 20:38:02 +0000 (20:38 +0000)]
less preprocessor magic in version number macros

Originally committed as revision 26105 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

16 years agoless preprocessor magic in version number macros
Måns Rullgård [Tue, 26 Feb 2008 20:37:59 +0000 (20:37 +0000)]
less preprocessor magic in version number macros

Originally committed as revision 12246 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago25% faster floating point AAN IDCT.
Michael Niedermayer [Tue, 26 Feb 2008 20:21:49 +0000 (20:21 +0000)]
25% faster floating point AAN IDCT.
Other float code likely benefits as well ...

Originally committed as revision 12245 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoStuffing bytes to not violate minimum legality rate.
Michael Niedermayer [Tue, 26 Feb 2008 20:17:30 +0000 (20:17 +0000)]
Stuffing bytes to not violate minimum legality rate.
Note, this is a joke and not related to MPEG CBR.

Originally committed as revision 12244 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoRemove another temporary variable with which gcc has problems.
Michael Niedermayer [Tue, 26 Feb 2008 18:36:58 +0000 (18:36 +0000)]
Remove another temporary variable with which gcc has problems.

Originally committed as revision 12243 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosupport keyframes indexes starting at 0, broken files created by omneon, fix ref...
Baptiste Coudurier [Tue, 26 Feb 2008 18:27:21 +0000 (18:27 +0000)]
support keyframes indexes starting at 0, broken files created by omneon, fix ref.mov seeking

Originally committed as revision 12242 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosupport mov reference files, ref.mov/ref.m2v/ref.wav and detect BrianCox.mov
Baptiste Coudurier [Tue, 26 Feb 2008 18:21:58 +0000 (18:21 +0000)]
support mov reference files, ref.mov/ref.m2v/ref.wav and detect BrianCox.mov

Originally committed as revision 12241 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetics: Demplayerify indentation.
Diego Biurrun [Tue, 26 Feb 2008 17:56:00 +0000 (17:56 +0000)]
cosmetics: Demplayerify indentation.

Originally committed as revision 12240 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAdd comments to some #endif directives.
Diego Biurrun [Tue, 26 Feb 2008 17:13:42 +0000 (17:13 +0000)]
Add comments to some #endif directives.

Originally committed as revision 12239 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetics: spelling/grammar
Diego Biurrun [Tue, 26 Feb 2008 17:06:31 +0000 (17:06 +0000)]
cosmetics: spelling/grammar

Originally committed as revision 12238 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAlternative and faster (on gcc x86) plane rotation.
Michael Niedermayer [Tue, 26 Feb 2008 15:52:37 +0000 (15:52 +0000)]
Alternative and faster (on gcc x86) plane rotation.

Originally committed as revision 12237 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoGet rid of another useless variable.
Michael Niedermayer [Tue, 26 Feb 2008 15:44:49 +0000 (15:44 +0000)]
Get rid of another useless variable.

Originally committed as revision 12236 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAvoid a=b+c but rather use a+=b (gcc is too stupid to do this itself),
Michael Niedermayer [Tue, 26 Feb 2008 15:32:34 +0000 (15:32 +0000)]
Avoid a=b+c but rather use a+=b (gcc is too stupid to do this itself),
code is exactly the same just uses different variables.

Originally committed as revision 12235 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoOne variable less. No real change to the actual code.
Michael Niedermayer [Tue, 26 Feb 2008 15:19:37 +0000 (15:19 +0000)]
One variable less. No real change to the actual code.

Originally committed as revision 12234 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agofloating point AAN IDCT
Michael Niedermayer [Tue, 26 Feb 2008 14:59:11 +0000 (14:59 +0000)]
floating point AAN IDCT

Originally committed as revision 12233 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoremove useless braces
Baptiste Coudurier [Tue, 26 Feb 2008 14:42:07 +0000 (14:42 +0000)]
remove useless braces

Originally committed as revision 12232 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoindentation
Baptiste Coudurier [Tue, 26 Feb 2008 14:41:40 +0000 (14:41 +0000)]
indentation

Originally committed as revision 12231 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agorescale frame size, fix i.3g2
Baptiste Coudurier [Tue, 26 Feb 2008 14:36:49 +0000 (14:36 +0000)]
rescale frame size, fix i.3g2

Originally committed as revision 12230 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agomove parameters reset after setting stream parameters, will need those infos
Baptiste Coudurier [Tue, 26 Feb 2008 14:31:36 +0000 (14:31 +0000)]
move parameters reset after setting stream parameters, will need those infos

Originally committed as revision 12229 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agodecode mng color decorrelation
Loren Merritt [Tue, 26 Feb 2008 10:21:33 +0000 (10:21 +0000)]
decode mng color decorrelation

Originally committed as revision 12228 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoRemove START_TIMER and STOP_TIMER macros.
Diego Biurrun [Tue, 26 Feb 2008 09:35:21 +0000 (09:35 +0000)]
Remove START_TIMER and STOP_TIMER macros.

Originally committed as revision 12227 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoDisable vhook on Interix, patch by Pete Ross, pross xvid org.
Peter Ross [Tue, 26 Feb 2008 09:14:55 +0000 (09:14 +0000)]
Disable vhook on Interix, patch by Pete Ross, pross xvid org.

Originally committed as revision 12226 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoThis header just needs stdint.h, not inttypes.h, which is a superset of
Peter Ross [Tue, 26 Feb 2008 09:12:55 +0000 (09:12 +0000)]
This header just needs stdint.h, not inttypes.h, which is a superset of
stdint.h.  This has the sideeffect of fixing the build on Interix.
patch by Pete Ross, pross xvid org

Originally committed as revision 12225 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosimplify
Loren Merritt [Tue, 26 Feb 2008 07:35:56 +0000 (07:35 +0000)]
simplify

Originally committed as revision 12224 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosplit encoding part of dsputil_mmx into its own file
Aurelien Jacobs [Mon, 25 Feb 2008 23:14:22 +0000 (23:14 +0000)]
split encoding part of dsputil_mmx into its own file

Originally committed as revision 12223 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoRemove commented-out av_log calls.
Diego Biurrun [Mon, 25 Feb 2008 22:44:35 +0000 (22:44 +0000)]
Remove commented-out av_log calls.

Originally committed as revision 12222 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoFix trellis quant + AAN DCT.
Michael Niedermayer [Mon, 25 Feb 2008 22:43:42 +0000 (22:43 +0000)]
Fix trellis quant + AAN DCT.

Originally committed as revision 12221 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoMake diego happy.
Michael Niedermayer [Mon, 25 Feb 2008 22:32:52 +0000 (22:32 +0000)]
Make diego happy.

Originally committed as revision 12220 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocompile various fdct only when encoders are enabled
Aurelien Jacobs [Mon, 25 Feb 2008 21:43:19 +0000 (21:43 +0000)]
compile various fdct only when encoders are enabled

Originally committed as revision 12219 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoFix MSVC identification, patch by Eddie Pang.
Eddie Pang [Mon, 25 Feb 2008 18:32:55 +0000 (18:32 +0000)]
Fix MSVC identification, patch by Eddie Pang.

Originally committed as revision 12218 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoMake code C conformant.
Michael Niedermayer [Mon, 25 Feb 2008 18:25:49 +0000 (18:25 +0000)]
Make code C conformant.

Originally committed as revision 12217 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoClarify the documentation for 'ffmpeg -v'.
Stefano Sabatini [Mon, 25 Feb 2008 16:01:53 +0000 (16:01 +0000)]
Clarify the documentation for 'ffmpeg -v'.
patch by Stefano Sabatini, stefano.sabatini-lala poste it

Originally committed as revision 12216 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoImprove wording/grammar of Doxygen comments.
Diego Biurrun [Mon, 25 Feb 2008 15:36:10 +0000 (15:36 +0000)]
Improve wording/grammar of Doxygen comments.

Originally committed as revision 12215 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoClarify which BSD license.
Michael Niedermayer [Mon, 25 Feb 2008 14:59:04 +0000 (14:59 +0000)]
Clarify which BSD license.

Originally committed as revision 12214 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetics: prettyprint
Diego Biurrun [Mon, 25 Feb 2008 14:53:55 +0000 (14:53 +0000)]
cosmetics: prettyprint

Originally committed as revision 12213 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAllow use of my AAN DCT under BSD, AMD has agreed to write and contribute
Michael Niedermayer [Mon, 25 Feb 2008 14:42:12 +0000 (14:42 +0000)]
Allow use of my AAN DCT under BSD, AMD has agreed to write and contribute
SSE(float) AAN DCT and IDCT to ffmpeg (no C++, no crap).

Originally committed as revision 12212 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoMake parse_options() use parse_number_or_die().
Stefano Sabatini [Mon, 25 Feb 2008 14:15:52 +0000 (14:15 +0000)]
Make parse_options() use parse_number_or_die().
Patch by Stefano Sabatini stefano sabatini-lala poste it

Originally committed as revision 12211 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoExport and document in cmdutils.h the function parse_number_or_die().
Stefano Sabatini [Mon, 25 Feb 2008 14:05:52 +0000 (14:05 +0000)]
Export and document in cmdutils.h the function parse_number_or_die().
Patch by Stefano Sabatini: stefano sabatini-lala poste it

Originally committed as revision 12210 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agofix mace duration
Baptiste Coudurier [Mon, 25 Feb 2008 12:59:42 +0000 (12:59 +0000)]
fix mace duration

Originally committed as revision 12209 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoprotect malloc overflow
Baptiste Coudurier [Mon, 25 Feb 2008 12:45:59 +0000 (12:45 +0000)]
protect malloc overflow

Originally committed as revision 12208 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoindentation
Baptiste Coudurier [Mon, 25 Feb 2008 12:44:15 +0000 (12:44 +0000)]
indentation

Originally committed as revision 12207 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agofail if extradata cannot be allocated
Baptiste Coudurier [Mon, 25 Feb 2008 12:42:32 +0000 (12:42 +0000)]
fail if extradata cannot be allocated

Originally committed as revision 12206 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago10l, protect malloc overflow
Baptiste Coudurier [Mon, 25 Feb 2008 12:38:37 +0000 (12:38 +0000)]
10l, protect malloc overflow

Originally committed as revision 12205 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosupport QDM2 in aiff, QDM2.aif
Baptiste Coudurier [Mon, 25 Feb 2008 12:29:34 +0000 (12:29 +0000)]
support QDM2 in aiff, QDM2.aif

Originally committed as revision 12204 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoremove useless close function
Baptiste Coudurier [Mon, 25 Feb 2008 12:24:30 +0000 (12:24 +0000)]
remove useless close function

Originally committed as revision 12203 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoindentation
Baptiste Coudurier [Mon, 25 Feb 2008 12:23:27 +0000 (12:23 +0000)]
indentation

Originally committed as revision 12202 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoremove useless casts
Baptiste Coudurier [Mon, 25 Feb 2008 12:15:26 +0000 (12:15 +0000)]
remove useless casts

Originally committed as revision 12201 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetics
Baptiste Coudurier [Mon, 25 Feb 2008 12:14:15 +0000 (12:14 +0000)]
cosmetics

Originally committed as revision 12200 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoremove useless tags, codec_get_id matches independently and qt uses lower case tags
Baptiste Coudurier [Mon, 25 Feb 2008 12:08:26 +0000 (12:08 +0000)]
remove useless tags, codec_get_id matches independently and qt uses lower case tags

Originally committed as revision 12199 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoset bps to uncompressed original sound data for compressed audio
Baptiste Coudurier [Mon, 25 Feb 2008 12:00:31 +0000 (12:00 +0000)]
set bps to uncompressed original sound data for compressed audio
according to aiff specs, qt set it to 16 for mace and ima4,
fail if block align is not set.

Originally committed as revision 12198 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agofix duration for ima4
Baptiste Coudurier [Mon, 25 Feb 2008 11:34:26 +0000 (11:34 +0000)]
fix duration for ima4

Originally committed as revision 12197 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoima4 in aiff support
Baptiste Coudurier [Mon, 25 Feb 2008 10:57:22 +0000 (10:57 +0000)]
ima4 in aiff support

Originally committed as revision 12196 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agopng filters
Loren Merritt [Mon, 25 Feb 2008 10:34:23 +0000 (10:34 +0000)]
png filters

Originally committed as revision 12195 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoFix outdated comment.
Diego Biurrun [Mon, 25 Feb 2008 09:43:52 +0000 (09:43 +0000)]
Fix outdated comment.

Originally committed as revision 26099 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

16 years ago#include "libavutil/avutil.h" in swscale.h
Måns Rullgård [Mon, 25 Feb 2008 09:24:33 +0000 (09:24 +0000)]
#include "libavutil/avutil.h" in swscale.h

Originally committed as revision 26097 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

16 years agoInstall headers in $prefix/include/$libname
Måns Rullgård [Mon, 25 Feb 2008 09:22:11 +0000 (09:22 +0000)]
Install headers in $prefix/include/$libname

Install each library's headers in $prefix/include/$libname.
The installed headers use #include "lib*/*.h"; applications
are expected to do the same.

Specify only $prefix/include in .pc files.

Originally committed as revision 12194 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoremove a memcpy
Loren Merritt [Mon, 25 Feb 2008 05:46:53 +0000 (05:46 +0000)]
remove a memcpy

Originally committed as revision 12193 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetics: Sort and prettyprint codec_wav_tags[].
Diego Biurrun [Sun, 24 Feb 2008 23:02:47 +0000 (23:02 +0000)]
cosmetics: Sort and prettyprint codec_wav_tags[].

Originally committed as revision 12192 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agofix memory leaks in vp3 decoder
Måns Rullgård [Sun, 24 Feb 2008 18:56:52 +0000 (18:56 +0000)]
fix memory leaks in vp3 decoder

Originally committed as revision 12191 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoSet indexmem to a saner default.
Michael Niedermayer [Sun, 24 Feb 2008 16:28:35 +0000 (16:28 +0000)]
Set indexmem to a saner default.

Originally committed as revision 12190 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago__asm __volatile -> asm volatile part 3
Reimar Döffinger [Sun, 24 Feb 2008 14:59:16 +0000 (14:59 +0000)]
__asm __volatile -> asm volatile part 3

Originally committed as revision 26091 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

16 years ago__asm __volatile -> asm volatile part 2
Reimar Döffinger [Sun, 24 Feb 2008 14:47:42 +0000 (14:47 +0000)]
__asm __volatile -> asm volatile part 2

Originally committed as revision 12189 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago__asm __volatile -> asm volatile, improves code consistency and works
Reimar Döffinger [Sun, 24 Feb 2008 14:46:22 +0000 (14:46 +0000)]
__asm __volatile -> asm volatile, improves code consistency and works
(as far as that is possible) with the Sun C compiler.

Originally committed as revision 12188 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoFix bug I introduced in r11962.
Vitor Sessak [Sun, 24 Feb 2008 07:47:02 +0000 (07:47 +0000)]
Fix bug I introduced in r11962.
Thanks to Uoti Urpala for spotting it and proposing the
fix.

Originally committed as revision 12187 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoReduce initial fifo size, so as not to senselessly waste memory.
Michael Niedermayer [Sun, 24 Feb 2008 01:56:27 +0000 (01:56 +0000)]
Reduce initial fifo size, so as not to senselessly waste memory.

Originally committed as revision 12186 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agorealloc fifo
Michael Niedermayer [Sun, 24 Feb 2008 01:55:34 +0000 (01:55 +0000)]
realloc fifo

Originally committed as revision 12185 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoFix issue357
Michael Niedermayer [Sun, 24 Feb 2008 01:04:00 +0000 (01:04 +0000)]
Fix issue357
Do what the spec says, insane or not:
"
Format 0 (uncompressed) and Format 3 (uncompressed little-endian) are similar. Both encode
uncompressed audio samples. For 8-bit samples, the two formats are identical. For 16-bit
samples, the two formats differ in byte ordering. In Format 0, 16-bit samples are encoded and
decoded according to the native byte ordering of the platform on which the encoder and Flash
Player, respectively, are running. In Format 3, 16-bit samples are always encoded in little-endian
order (least significant byte first), and are byte-swapped if necessary in Flash Player before
playback. Format 0 is clearly disadvantageous because it introduces a playback platform
dependency. For 16-bit samples, Format 3 is highly preferable to Format 0 for SWF version 4
or later.
"

Originally committed as revision 12184 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoflv/swf do not have a big endian codec id, they only support
Michael Niedermayer [Sun, 24 Feb 2008 00:57:15 +0000 (00:57 +0000)]
flv/swf do not have a big endian codec id, they only support
little endian and native endian.

Originally committed as revision 12183 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoRiff ids for wmapro and wmavoice, this adds support for stream copy of these codecs...
Benjamin Larsson [Sat, 23 Feb 2008 23:30:41 +0000 (23:30 +0000)]
Riff ids for wmapro and wmavoice, this adds support for stream copy of these codecs between asf,avi and wav containers.

Originally committed as revision 12182 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agouse diff -u -w in codec regression test
Måns Rullgård [Sat, 23 Feb 2008 13:15:24 +0000 (13:15 +0000)]
use diff -u -w in codec regression test

Originally committed as revision 12181 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agotreat stream_type==0x82 as AUDIO_DTS when the program has a REGISTRATION_DESCRIPTOR...
Nico Sabbi [Sat, 23 Feb 2008 10:57:38 +0000 (10:57 +0000)]
treat stream_type==0x82 as AUDIO_DTS when the program has a REGISTRATION_DESCRIPTOR with the value HDMV; approved by Mans

Originally committed as revision 12180 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoRemove apparently unneeded and buggy align.
Michael Niedermayer [Fri, 22 Feb 2008 19:27:22 +0000 (19:27 +0000)]
Remove apparently unneeded and buggy align.

Originally committed as revision 12179 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoAdd -lm to x264 check, it is needed on most systems.
Diego Biurrun [Fri, 22 Feb 2008 17:15:24 +0000 (17:15 +0000)]
Add -lm to x264 check, it is needed on most systems.
taken from the DOS/DJGPP port by Michael Kostylev, mik niipt ru

Originally committed as revision 12178 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocleaner way of writing packed-byte constants.
Loren Merritt [Fri, 22 Feb 2008 16:53:26 +0000 (16:53 +0000)]
cleaner way of writing packed-byte constants.
suggested by Uoti Urpala

Originally committed as revision 12177 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoDiscard two symbols from libswscale.
Carl Eugen Hoyos [Fri, 22 Feb 2008 08:08:19 +0000 (08:08 +0000)]
Discard two symbols from libswscale.

Originally committed as revision 26060 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

16 years agoAscii art to explain what avfilter_formats_changeref() does.
Michael Niedermayer [Thu, 21 Feb 2008 22:33:40 +0000 (22:33 +0000)]
Ascii art to explain what avfilter_formats_changeref() does.

Originally committed as revision 12176 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agocosmetic
Michael Niedermayer [Thu, 21 Feb 2008 22:26:24 +0000 (22:26 +0000)]
cosmetic

Originally committed as revision 12175 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoBetter description (merge is too vague IMHO).
Michael Niedermayer [Thu, 21 Feb 2008 22:19:51 +0000 (22:19 +0000)]
Better description (merge is too vague IMHO).

Originally committed as revision 12174 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoBetter description for merge_ref().
Michael Niedermayer [Thu, 21 Feb 2008 22:10:11 +0000 (22:10 +0000)]
Better description for merge_ref().

Originally committed as revision 12173 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoDocument avfilter_formats_ref() with some ascii art.
Michael Niedermayer [Thu, 21 Feb 2008 22:02:53 +0000 (22:02 +0000)]
Document avfilter_formats_ref() with some ascii art.

Originally committed as revision 12172 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agominor English corrections
Mike Melanson [Thu, 21 Feb 2008 18:46:49 +0000 (18:46 +0000)]
minor English corrections

Originally committed as revision 12171 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agotypos
Michael Niedermayer [Thu, 21 Feb 2008 18:42:33 +0000 (18:42 +0000)]
typos

Originally committed as revision 12170 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoExplain that libavcodec/i386/idct_mmx.c is GPLed.
Diego Biurrun [Thu, 21 Feb 2008 15:48:58 +0000 (15:48 +0000)]
Explain that libavcodec/i386/idct_mmx.c is GPLed.

Originally committed as revision 12169 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years ago"General Tips" section
Michael Niedermayer [Thu, 21 Feb 2008 13:02:31 +0000 (13:02 +0000)]
"General Tips" section

Originally committed as revision 12168 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoparse_number_or_die()
Michael Niedermayer [Thu, 21 Feb 2008 12:24:37 +0000 (12:24 +0000)]
parse_number_or_die()
Based on a patch by Stefano Sabatini.

Originally committed as revision 12167 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agoindent
Loren Merritt [Thu, 21 Feb 2008 07:55:42 +0000 (07:55 +0000)]
indent

Originally committed as revision 12166 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agopseudo-simd add_bytes and diff_bytes
Loren Merritt [Thu, 21 Feb 2008 07:54:46 +0000 (07:54 +0000)]
pseudo-simd add_bytes and diff_bytes
2x faster than scalar in 32bit, 4x faster in 64bit (as opposed to 8x in mmx)

Originally committed as revision 12165 to svn://svn.ffmpeg.org/ffmpeg/trunk

16 years agosimd and unroll png_filter_row
Loren Merritt [Thu, 21 Feb 2008 07:10:46 +0000 (07:10 +0000)]
simd and unroll png_filter_row
cycles per 1000 pixels on core2:
left: 9211->5170
top: 9283->2138
avg: 12215->7611
paeth: 64024->17360
overall rgb png decoding speed: +45%
overall greyscale png decoding speed: +6%

Originally committed as revision 12164 to svn://svn.ffmpeg.org/ffmpeg/trunk