ffmpeg.git
10 years agoChange the order of parameters for ff_eval_expr() and
Stefano Sabatini [Sun, 16 May 2010 23:00:22 +0000 (23:00 +0000)]
Change the order of parameters for ff_eval_expr() and
ff_parse_and_eval_expr(), place the names for constants/functions
before the corresponding values.

This looks more readable, as the user is expected to know the names
before the values.

Originally committed as revision 23149 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFactorize code into a single function.
Sebastian Vater [Sun, 16 May 2010 21:03:08 +0000 (21:03 +0000)]
Factorize code into a single function.

Patch by Sebastian Vater, cdgs D basty A gmail

Originally committed as revision 23148 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agouse filter name when graph parser add filters
Baptiste Coudurier [Sun, 16 May 2010 18:13:55 +0000 (18:13 +0000)]
use filter name when graph parser add filters

Originally committed as revision 23147 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd missing rawvideo pixel formats to codec tags mappings for nut.
Stefano Sabatini [Sun, 16 May 2010 17:08:47 +0000 (17:08 +0000)]
Add missing rawvideo pixel formats to codec tags mappings for nut.

Originally committed as revision 23146 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd NV12 and NV21 AVI tags.
Stefano Sabatini [Sun, 16 May 2010 07:56:31 +0000 (07:56 +0000)]
Add NV12 and NV21 AVI tags.

Both are listed in fourcc.org.

Originally committed as revision 23145 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoClarify description for the MONOWHITE and MONOBLACK pixel formats.
Stefano Sabatini [Sun, 16 May 2010 07:56:26 +0000 (07:56 +0000)]
Clarify description for the MONOWHITE and MONOBLACK pixel formats.

Originally committed as revision 23144 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAvoid mixed declaration and code, fix C89 compatibility.
François Revol [Sat, 15 May 2010 17:34:45 +0000 (17:34 +0000)]
Avoid mixed declaration and code, fix C89 compatibility.

Patch by François Revol revol free fr.

Originally committed as revision 23143 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoClarify descriptions for RGB4, BGR4, NV12, NV21,
Stefano Sabatini [Sat, 15 May 2010 12:33:31 +0000 (12:33 +0000)]
Clarify descriptions for RGB4, BGR4, NV12, NV21,
RGB48BE, and RGB48LE pixel formats.

Originally committed as revision 23142 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReindent after last commit on dxva2 h264 AVHWAccel.
Laurent Aimar [Fri, 14 May 2010 23:09:08 +0000 (23:09 +0000)]
Reindent after last commit on dxva2 h264 AVHWAccel.

Originally committed as revision 23141 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFixed h264 long term support with dxva2 AVHWAccel.
Laurent Aimar [Fri, 14 May 2010 23:08:23 +0000 (23:08 +0000)]
Fixed h264 long term support with dxva2 AVHWAccel.

Based on a commit for vaapi(r22869).

Originally committed as revision 23140 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoChange MAX_READ_SIZE message during av_find_stream_info to DEBUG level.
Baptiste Coudurier [Fri, 14 May 2010 22:04:39 +0000 (22:04 +0000)]
Change MAX_READ_SIZE message during av_find_stream_info to DEBUG level.
It is not harmful and it scares too many users.

Originally committed as revision 23139 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agofmaxf -> FFMAX to fix pre-C99 systems
Alex Converse [Fri, 14 May 2010 19:38:17 +0000 (19:38 +0000)]
fmaxf -> FFMAX to fix pre-C99 systems

Originally committed as revision 23138 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoChange inter-protocol dependencies from _deps to _select
Martin Storsjö [Fri, 14 May 2010 18:51:46 +0000 (18:51 +0000)]
Change inter-protocol dependencies from _deps to _select

Originally committed as revision 23137 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoaacenc: Add a rate only trellis for codebook selection for the TLS.
Alex Converse [Fri, 14 May 2010 16:49:51 +0000 (16:49 +0000)]
aacenc: Add a rate only trellis for codebook selection for the TLS.

Originally committed as revision 23136 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoaacenc: Use exact values when quantizing, not fuzzy values.
Alex Converse [Fri, 14 May 2010 16:49:40 +0000 (16:49 +0000)]
aacenc: Use exact values when quantizing, not fuzzy values.

This requires us to code small escapes; we can't avoid it.

Originally committed as revision 23135 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoaacenc: Use an estimated codebook for the TLS (two loop search).
Alex Converse [Fri, 14 May 2010 16:49:30 +0000 (16:49 +0000)]
aacenc: Use an estimated codebook for the TLS (two loop search).

Originally committed as revision 23134 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoaacenc: Select the TLS (two-loop search) as the default scalefactor coder.
Alex Converse [Fri, 14 May 2010 16:49:21 +0000 (16:49 +0000)]
aacenc: Select the TLS (two-loop search) as the default scalefactor coder.

Originally committed as revision 23133 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoaacenc: Fix psy logic.
Alex Converse [Fri, 14 May 2010 16:49:11 +0000 (16:49 +0000)]
aacenc: Fix psy logic.

Set band info before determining scalefactors. Use the look ahead for
windowing decision.

Originally committed as revision 23132 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFix compilation of AC3 decoder if E-AC3 decoder was disabled.
Carl Eugen Hoyos [Thu, 13 May 2010 20:41:36 +0000 (20:41 +0000)]
Fix compilation of AC3 decoder if E-AC3 decoder was disabled.

Originally committed as revision 23131 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoqt-faststart: Abort scanning of the input file if a badly sized atom is encountered
Martin Storsjö [Thu, 13 May 2010 19:22:56 +0000 (19:22 +0000)]
qt-faststart: Abort scanning of the input file if a badly sized atom is encountered

If the atom size is 0, qt-faststart currently hangs forever while scanning
the file.

Originally committed as revision 23129 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReindent after r23124. Patch by Sebastian Vater <cdgs basty googlemail com>.
Sebastian Vater [Thu, 13 May 2010 19:18:54 +0000 (19:18 +0000)]
Reindent after r23124. Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23128 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoCosmetics: Initialize pointers with NULL instead of 0, for consistency
Martin Storsjö [Thu, 13 May 2010 19:18:27 +0000 (19:18 +0000)]
Cosmetics: Initialize pointers with NULL instead of 0, for consistency

Originally committed as revision 23127 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoCosmetics: reindent
Martin Storsjö [Thu, 13 May 2010 19:17:49 +0000 (19:17 +0000)]
Cosmetics: reindent

Originally committed as revision 23126 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoqt-faststart: Use the error_out cleanup code path for all error returns
Martin Storsjö [Thu, 13 May 2010 19:17:15 +0000 (19:17 +0000)]
qt-faststart: Use the error_out cleanup code path for all error returns

Originally committed as revision 23125 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoGrayscale support. Patch by Sebastian Vater <cdgs basty googlemail com>.
Sebastian Vater [Thu, 13 May 2010 19:16:03 +0000 (19:16 +0000)]
Grayscale support. Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23124 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd mpegaudiodec_float.o dependency on tables header with hardcoded tables
Måns Rullgård [Thu, 13 May 2010 18:41:18 +0000 (18:41 +0000)]
Add mpegaudiodec_float.o dependency on tables header with hardcoded tables

Originally committed as revision 23123 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRemove the name of the file from the @file doxy, it is unnecessary and
Stefano Sabatini [Thu, 13 May 2010 18:26:11 +0000 (18:26 +0000)]
Remove the name of the file from the @file doxy, it is unnecessary and
inconsistent with the other files.

Originally committed as revision 23122 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake the init and config_filter callbacks of the pad filter return
Stefano Sabatini [Thu, 13 May 2010 18:26:07 +0000 (18:26 +0000)]
Make the init and config_filter callbacks of the pad filter return
AVERROR(EINVAL) rather than -1 in case of invalid parameters.

Originally committed as revision 23121 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake av_parse_color() return AVERROR(EINVAL) rather than -1.
Stefano Sabatini [Thu, 13 May 2010 18:26:03 +0000 (18:26 +0000)]
Make av_parse_color() return AVERROR(EINVAL) rather than -1.

Originally committed as revision 23120 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years ago1.0 and the resulting exactly representable value must be marked as float as well,
Michael Niedermayer [Thu, 13 May 2010 18:06:44 +0000 (18:06 +0000)]
1.0 and the resulting exactly representable value must be marked as float as well,
gcc is hopelessly trash.

Originally committed as revision 23119 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoCast constants to float to avoid gcc converting to and from
Michael Niedermayer [Thu, 13 May 2010 17:58:17 +0000 (17:58 +0000)]
Cast constants to float to avoid gcc converting to and from
float<->double in every operation.

Originally committed as revision 23118 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoUse standard C for implementing sign_extend() and zero_extend().
Michael Niedermayer [Thu, 13 May 2010 16:32:21 +0000 (16:32 +0000)]
Use standard C for implementing sign_extend() and zero_extend().
This fixes compilation of probetest

Originally committed as revision 23116 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoav_alias is an attribute and belongs to attributes.h
Michael Niedermayer [Thu, 13 May 2010 16:30:58 +0000 (16:30 +0000)]
av_alias is an attribute and belongs to attributes.h
also attributes.h is public and external api and can thus not depend
on configure tested compiler support thus this part is removed. A
different solution must be found if this breaks for some compiler
which i hope it does not.

Originally committed as revision 23115 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReindent after r23112.
Sebastian Vater [Thu, 13 May 2010 15:41:49 +0000 (15:41 +0000)]
Reindent after r23112.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23113 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMove handling of paletted data to the IFF demuxer. This allows future
Sebastian Vater [Thu, 13 May 2010 15:39:40 +0000 (15:39 +0000)]
Move handling of paletted data to the IFF demuxer. This allows future
handling of things such as masking/EHB/HAM for this type of data.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23112 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoHandle palette underflows, fill remaining space with black (zero) data.
Sebastian Vater [Thu, 13 May 2010 15:33:36 +0000 (15:33 +0000)]
Handle palette underflows, fill remaining space with black (zero) data.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23111 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMove a while(..){..} -> do{..}while(..), slightly faster.
Sebastian Vater [Thu, 13 May 2010 13:18:25 +0000 (13:18 +0000)]
Move a while(..){..} -> do{..}while(..), slightly faster.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23110 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReplace "-vfilters" by "-vf" in regtests. Should fix regtest breakage.
Vitor Sessak [Thu, 13 May 2010 12:30:46 +0000 (12:30 +0000)]
Replace "-vfilters" by "-vf" in regtests. Should fix regtest breakage.

Originally committed as revision 23109 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFATE: change -vfilters to -vf
Måns Rullgård [Thu, 13 May 2010 11:32:13 +0000 (11:32 +0000)]
FATE: change -vfilters to -vf

Originally committed as revision 23108 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFix compilation with low precission mpeg audio decoding.
Michael Niedermayer [Thu, 13 May 2010 10:04:31 +0000 (10:04 +0000)]
Fix compilation with low precission mpeg audio decoding.

Originally committed as revision 23107 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFix mpegaudio tablegen.
Michael Kostylev [Thu, 13 May 2010 10:02:48 +0000 (10:02 +0000)]
Fix mpegaudio tablegen.

Patch by Michael Kostylev, michael D kostylev A gmail

Originally committed as revision 23106 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoSupport writing 2d float arrays.
Michael Kostylev [Thu, 13 May 2010 10:02:04 +0000 (10:02 +0000)]
Support writing 2d float arrays.

Patch by Michael Kostylev, michael D kostylev A gmail

Originally committed as revision 23105 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agorename -vfilters cli option to -vf in ffplay as well
Baptiste Coudurier [Wed, 12 May 2010 23:18:12 +0000 (23:18 +0000)]
rename -vfilters cli option to -vf in ffplay as well

Originally committed as revision 23104 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agochange a variable-length array to a malloc.
Loren Merritt [Wed, 12 May 2010 22:38:05 +0000 (22:38 +0000)]
change a variable-length array to a malloc.

Originally committed as revision 23103 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoDoxygen av_codec_get_id() and av_codec_get_tag().
Stefano Sabatini [Wed, 12 May 2010 22:17:29 +0000 (22:17 +0000)]
Doxygen av_codec_get_id() and av_codec_get_tag().

Originally committed as revision 23102 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRevert r31153. It failed to build on:
Ramiro Polla [Wed, 12 May 2010 20:47:32 +0000 (20:47 +0000)]
Revert r31153. It failed to build on:
x86_64 / Mac OS X gcc 4.0.1
x86_64 / Linux icc (all)
x86_64 / Linux gcc 4.0.4
x86_64 / OpenBSD gcc 3.3.5
x86_64 / Linux suncc 5.10
and there are some reports of crashes.

Originally committed as revision 31170 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

10 years agoSet coded_frame to NULL when closing a codec, since it might
Reimar Döffinger [Wed, 12 May 2010 17:05:56 +0000 (17:05 +0000)]
Set coded_frame to NULL when closing a codec, since it might
be invalid after the codec is "gone".

Originally committed as revision 23101 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agorename -vfilters cli option to -vf
Baptiste Coudurier [Wed, 12 May 2010 01:47:14 +0000 (01:47 +0000)]
rename -vfilters cli option to -vf

Originally committed as revision 23100 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoDo the same sign flip optimization to the low freq decoder.
Michael Niedermayer [Tue, 11 May 2010 21:31:54 +0000 (21:31 +0000)]
Do the same sign flip optimization to the low freq decoder.
as with the high freq 10-20 cycles faster

Originally committed as revision 23099 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFactorize READ_FLIP_SIGN() optimization out
Michael Niedermayer [Tue, 11 May 2010 21:20:33 +0000 (21:20 +0000)]
Factorize READ_FLIP_SIGN() optimization out

Originally committed as revision 23098 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoOptimize decoding high freqs.
Michael Niedermayer [Tue, 11 May 2010 21:10:55 +0000 (21:10 +0000)]
Optimize decoding high freqs.
this is 10-20cpu cycles faster on duron (whole is about 50-60 cpu cylses)
I wonder why gcc isnt doing this on its own ...

Originally committed as revision 23097 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake lsf_sf_expand() 4 times faster.
Michael Niedermayer [Tue, 11 May 2010 20:33:48 +0000 (20:33 +0000)]
Make lsf_sf_expand() 4 times faster.

Originally committed as revision 23096 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agofloat based mp1/mp2/mp3 decoders.
Michael Niedermayer [Tue, 11 May 2010 19:52:42 +0000 (19:52 +0000)]
float based mp1/mp2/mp3 decoders.

Originally committed as revision 23095 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoSet cur_channel in the AAC encoder context where needed.
Alex Converse [Tue, 11 May 2010 18:53:49 +0000 (18:53 +0000)]
Set cur_channel in the AAC encoder context where needed.

Most coder functions read it. Carting this around in the context may be
suboptimal; a refactor should be considered.

Originally committed as revision 23094 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agosilence gcc warning about potential uninitialized usage
Baptiste Coudurier [Tue, 11 May 2010 17:39:52 +0000 (17:39 +0000)]
silence gcc warning about potential uninitialized usage

Originally committed as revision 23093 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agocosmetics, rename loop to frame_available
Baptiste Coudurier [Tue, 11 May 2010 17:25:18 +0000 (17:25 +0000)]
cosmetics, rename loop to frame_available

Originally committed as revision 23092 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agorename curr_filter to last_filter, factorize filter declaration
Baptiste Coudurier [Tue, 11 May 2010 17:23:06 +0000 (17:23 +0000)]
rename curr_filter to last_filter, factorize filter declaration

Originally committed as revision 23091 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agocosmetics: indentation, whitespaces
Baptiste Coudurier [Tue, 11 May 2010 17:18:13 +0000 (17:18 +0000)]
cosmetics: indentation, whitespaces

Originally committed as revision 23090 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agosimplify, reuse existing args variable
Baptiste Coudurier [Tue, 11 May 2010 17:16:52 +0000 (17:16 +0000)]
simplify, reuse existing args variable

Originally committed as revision 23089 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agocosmetics: filt_graph_all -> graph, like in ffplay.c
Baptiste Coudurier [Tue, 11 May 2010 17:14:36 +0000 (17:14 +0000)]
cosmetics: filt_graph_all -> graph, like in ffplay.c

Originally committed as revision 23088 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd -ldl to libfaadbin_extralibs instead of libfaadbin_decoder_extralibs.
Janne Grunau [Tue, 11 May 2010 15:58:50 +0000 (15:58 +0000)]
Add -ldl to libfaadbin_extralibs instead of libfaadbin_decoder_extralibs.
The latter does not exist and thus compilation fails.
patch by Janne Grunau, janne-ffmpeg jannau.net

Originally committed as revision 23087 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRemove unused FRAC_RND() macro from mpegaudiodec.c.
Michael Niedermayer [Tue, 11 May 2010 14:49:48 +0000 (14:49 +0000)]
Remove unused FRAC_RND() macro from mpegaudiodec.c.

Originally committed as revision 23086 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agomlpdec: Comment channel_params field in struct SubStream.
Nick Brereton [Tue, 11 May 2010 01:46:13 +0000 (01:46 +0000)]
mlpdec: Comment channel_params field in struct SubStream.
Patch by Nick Brereton <nick at nbrereton dot net>

Originally committed as revision 23085 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agomlpdec: Allocate channel decoding parameters for each substream. Some file
Nick Brereton [Tue, 11 May 2010 01:44:52 +0000 (01:44 +0000)]
mlpdec: Allocate channel decoding parameters for each substream. Some file
was encountered with a channel range that overlapped the previous substreams,
and the code assumed no such overlap was possible.
Patch by Nick Brereton <nick at nbrereton dot net>

Originally committed as revision 23084 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoUse int instead of long to pass width parameters in non-public functions.
Ramiro Polla [Tue, 11 May 2010 00:36:34 +0000 (00:36 +0000)]
Use int instead of long to pass width parameters in non-public functions.
long was being incorrectly used as an x86-sized register, both for 32 and 64
bits, but this is not the case in win64.

Originally committed as revision 31153 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

10 years agoAdding missing () to mathops.h.
Michael Niedermayer [Tue, 11 May 2010 00:22:50 +0000 (00:22 +0000)]
Adding missing () to mathops.h.

Originally committed as revision 23083 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReplace more "m" constraints with MANGLE to fix compilation issues
Reimar Döffinger [Mon, 10 May 2010 21:16:08 +0000 (21:16 +0000)]
Replace more "m" constraints with MANGLE to fix compilation issues
with x86_32 gcc 4.4.4 and -fPIC.

Originally committed as revision 23082 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoOptimize decodeplane32().
Sebastian Vater [Mon, 10 May 2010 17:18:09 +0000 (17:18 +0000)]
Optimize decodeplane32().

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23081 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoEnsure that width and height are > 0. avcodec_open() itself only checks that
Sebastian Vater [Mon, 10 May 2010 17:00:56 +0000 (17:00 +0000)]
Ensure that width and height are > 0. avcodec_open() itself only checks that
they are >= 0.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23080 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFATE: print friendly error for individual tests when SAMPLES unset
Måns Rullgård [Mon, 10 May 2010 09:32:59 +0000 (09:32 +0000)]
FATE: print friendly error for individual tests when SAMPLES unset

Originally committed as revision 23079 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoCheck NAL unit size to avoid reading past the buffer.
Thomas Devanneaux [Mon, 10 May 2010 07:08:57 +0000 (07:08 +0000)]
Check NAL unit size to avoid reading past the buffer.
This fixes issue1907

Patch by Thomas Devanneaux gmail(thomdev)

Originally committed as revision 23078 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAlloc 16 extra bytes in libavfilter frames. Needed for MMX-optimized swscale.
Vitor Sessak [Mon, 10 May 2010 00:28:18 +0000 (00:28 +0000)]
Alloc 16 extra bytes in libavfilter frames. Needed for MMX-optimized swscale.

Fix issue 1924.

Originally committed as revision 23077 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd FourCC MJPG for CODEC_ID_JPEGLS.
Francesco Lavra [Sun, 9 May 2010 23:19:19 +0000 (23:19 +0000)]
Add FourCC MJPG for CODEC_ID_JPEGLS.

Patch by Francesco Lavra, francescolavra interfree it

Originally committed as revision 23076 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReorder nut specific codec tags and add a comment for marking them as
Stefano Sabatini [Sun, 9 May 2010 23:04:06 +0000 (23:04 +0000)]
Reorder nut specific codec tags and add a comment for marking them as
such.

Also put the [3][0][0][0] codec tag, mapped to rgb565le, in a special
section. It needs to be specified *after* the nut RGB[16] codec tag,
otherwise it will be used by default when encoding normal non-flipped
rgb565le, and will be decoded like a flipped format (see
rawdec.c:raw_init_decoder()).

Originally committed as revision 23075 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd missing nut-specific codec tags for rawvideo pixel formats.
Stefano Sabatini [Sun, 9 May 2010 23:04:03 +0000 (23:04 +0000)]
Add missing nut-specific codec tags for rawvideo pixel formats.

Add codec tags for the formats:
[15]BGR Packed RGB 5:5:5, 16bpp, (msb)1A 5R 5G 5B(lsb), big-endian [NOT in AVI]
[15]RGB Packed BGR 5:5:5, 16bpp, (msb)1A 5B 5G 5R(lsb), big-endian [NOT in AVI]
[16]BGR Packed RGB 5:6:5, 16bpp, (msb)   5R 6G 5B(lsb), big-endian [NOT in AVI]
[16]RGB Packed BGR 5:6:5, 16bpp, (msb)   5B 6G 5R(lsb), big-endian [NOT in AVI]
RGBA    Packed RGBA 8:8:8:8, 32bpp, RGBA [NOT in AVI]
BGRA    Packed RGBA 8:8:8:8, 32bpp, BGRA [NOT in AVI]
ABGR    Packed RGBA 8:8:8:8, 32bpp, ABGR [NOT in AVI]
ARGB    Packed RGBA 8:8:8:8, 32bpp, ARGB [NOT in AVI]
RGB[24] Packed RGB 8:8:8, 24bpp RGB [NOT in AVI]
BGR[24] Packed RGB 8:8:8, 24bpp BGR [NOT in AVI]

which are listed in the nut4cc.txt spec.

Originally committed as revision 23074 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake the codec tags for the yuvjXXX pixel formats the same as the
Stefano Sabatini [Sun, 9 May 2010 23:04:01 +0000 (23:04 +0000)]
Make the codec tags for the yuvjXXX pixel formats the same as the
corresponding ones for the yuvXXX pixel formats.

Originally committed as revision 23073 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd support to the Y411 codec tag, corresponding to the rawvideo pixel
Stefano Sabatini [Sun, 9 May 2010 22:56:04 +0000 (22:56 +0000)]
Add support to the Y411 codec tag, corresponding to the rawvideo pixel
format uyyvyy411.

The codec tag is referenced in fourcc.org.

Originally committed as revision 23072 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake the nut demuxer issue a more meaningful error message if it
Stefano Sabatini [Sun, 9 May 2010 22:56:02 +0000 (22:56 +0000)]
Make the nut demuxer issue a more meaningful error message if it
cannot recognize the provided codec tag.

Originally committed as revision 23071 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agobswap: 10L add missing parens around macro args
Måns Rullgård [Sun, 9 May 2010 20:32:35 +0000 (20:32 +0000)]
bswap: 10L add missing parens around macro args

Originally committed as revision 23070 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoReplace usage of s->streams[0]->* with st->*, which is shorter.
Sebastian Vater [Sun, 9 May 2010 20:10:48 +0000 (20:10 +0000)]
Replace usage of s->streams[0]->* with st->*, which is shorter.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23069 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRemove "bps" parameter to decodeplane8/32(), it's unused.
Sebastian Vater [Sun, 9 May 2010 20:08:13 +0000 (20:08 +0000)]
Remove "bps" parameter to decodeplane8/32(), it's unused.

Patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23068 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd entry for AVFilterPicRef interlaced and top_field_first fields
Stefano Sabatini [Sun, 9 May 2010 20:07:29 +0000 (20:07 +0000)]
Add entry for AVFilterPicRef interlaced and top_field_first fields
addition.

Originally committed as revision 23067 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoBump lavfi minor after the addition of the fields interlaced and
Stefano Sabatini [Sun, 9 May 2010 20:02:40 +0000 (20:02 +0000)]
Bump lavfi minor after the addition of the fields interlaced and
top_field_first in AVFilterPicRef, done in r23044.

Originally committed as revision 23066 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoqt-faststart: Avoid leaking memory if encountering a file with double ftyp atoms
Martin Storsjö [Sun, 9 May 2010 17:46:42 +0000 (17:46 +0000)]
qt-faststart: Avoid leaking memory if encountering a file with double ftyp atoms

Originally committed as revision 23065 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFATE: update idroq-video-encode command
Måns Rullgård [Sun, 9 May 2010 17:05:35 +0000 (17:05 +0000)]
FATE: update idroq-video-encode command

Originally committed as revision 23064 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoIFF: decode last 8 pixels per line
Måns Rullgård [Sun, 9 May 2010 16:24:20 +0000 (16:24 +0000)]
IFF: decode last 8 pixels per line

The decodeplane8() function processes one byte of input less than
it should.  Also, the for loop has an unusual style with side-effects
in the controlling expression; replaced with a more intuitive while
loop.

10l to Basty.

Originally committed as revision 23063 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFix build with swscale disabled
Måns Rullgård [Sun, 9 May 2010 12:04:48 +0000 (12:04 +0000)]
Fix build with swscale disabled

Originally committed as revision 23062 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoOptimize decodeplane8(), patch by Sebastian Vater <cdgs basty googlemail com>.
Sebastian Vater [Sat, 8 May 2010 21:48:44 +0000 (21:48 +0000)]
Optimize decodeplane8(), patch by Sebastian Vater <cdgs basty googlemail com>.

Originally committed as revision 23061 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoFix auto-scaling.
Stefano Sabatini [Sat, 8 May 2010 21:39:57 +0000 (21:39 +0000)]
Fix auto-scaling.

Use the numeric value assigned to sws_flags for the sws_flags set in
the graph, rather than the string "bilinear", which is not even
parsable by the scale filter.

Originally committed as revision 23060 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAdd means to adjust the log level per context.
Michael Niedermayer [Sat, 8 May 2010 20:41:20 +0000 (20:41 +0000)]
Add means to adjust the log level per context.

Originally committed as revision 23059 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRemove hardcoded-tables hack for IA-64: with latest binutils that now actually
Reimar Döffinger [Sat, 8 May 2010 16:32:05 +0000 (16:32 +0000)]
Remove hardcoded-tables hack for IA-64: with latest binutils that now actually
causes linking errors instead of avoiding them.

Originally committed as revision 23058 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoAnother try for fixing/improving decode_video documentation.
Reimar Döffinger [Sat, 8 May 2010 16:24:33 +0000 (16:24 +0000)]
Another try for fixing/improving decode_video documentation.

Originally committed as revision 23057 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoMake config_props() show conversion information before to create the
Stefano Sabatini [Fri, 7 May 2010 22:12:13 +0000 (22:12 +0000)]
Make config_props() show conversion information before to create the
swscale context.

This makes eventual warnings issued in case of swscale context
creation failure to be shown after the conversion information rather
than before, which is slightly less confusing.

Originally committed as revision 23056 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoLog input size, input format and swscale flags used for conversion in
Stefano Sabatini [Fri, 7 May 2010 22:12:10 +0000 (22:12 +0000)]
Log input size, input format and swscale flags used for conversion in
config_props().

Useful for debugging.

Originally committed as revision 23055 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoblackfin: fix yuv422 to yuv420 conversion
Ronaldo Moura [Fri, 7 May 2010 21:59:18 +0000 (21:59 +0000)]
blackfin: fix yuv422 to yuv420 conversion

The old code is correct only when stride = 2*width.

Patch by Ronaldo Moura <ronaldo d moura monity com br>

Originally committed as revision 31142 to svn://svn.mplayerhq.hu/mplayer/trunk/libswscale

10 years agovf_pad: fix mixed code and declarations
Måns Rullgård [Fri, 7 May 2010 17:53:29 +0000 (17:53 +0000)]
vf_pad: fix mixed code and declarations

Originally committed as revision 23054 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoc99 sucks. Replacing scanf("%i") by strtoul()
Michael Niedermayer [Fri, 7 May 2010 13:00:56 +0000 (13:00 +0000)]
c99 sucks. Replacing scanf("%i") by strtoul()

Originally committed as revision 23053 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoDocument the pad filter.
Stefano Sabatini [Fri, 7 May 2010 12:52:46 +0000 (12:52 +0000)]
Document the pad filter.

Originally committed as revision 23052 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoDocument cmdutils.c:print_error().
Stefano Sabatini [Fri, 7 May 2010 12:42:26 +0000 (12:42 +0000)]
Document cmdutils.c:print_error().

Originally committed as revision 23051 to svn://svn.ffmpeg.org/ffmpeg/trunk

10 years agoRemove messy pading hack in ffmpeg.c.
Michael Niedermayer [Fri, 7 May 2010 12:16:23 +0000 (12:16 +0000)]
Remove messy pading hack in ffmpeg.c.
Use avfilters if you want padding!

Originally committed as revision 23050 to svn://svn.ffmpeg.org/ffmpeg/trunk