ffmpeg.git
8 years agodrop support for base-10 encoded alpha component in av_parse_color()
Aurelien Jacobs [Wed, 10 Nov 2010 23:37:12 +0000 (23:37 +0000)]
drop support for base-10 encoded alpha component in av_parse_color()

Originally committed as revision 25719 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoIn avcodec_pix_fmt_string() use local variable pixdesc rather than
Stefano Sabatini [Wed, 10 Nov 2010 20:13:24 +0000 (20:13 +0000)]
In avcodec_pix_fmt_string() use local variable pixdesc rather than
av_pix_fmt_descriptors[pix_fmt], simplify.

Originally committed as revision 25718 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake avcodec_pix_fmt_string() use the information in the pixel format
Stefano Sabatini [Wed, 10 Nov 2010 20:13:19 +0000 (20:13 +0000)]
Make avcodec_pix_fmt_string() use the information in the pixel format
descriptors for printing the number of channels/components.

Also replace the term "nb_channels" with "nb_components" which is more
consistent with the FFmpeg internal terminology, and is somehow
different with respect to the current definition of nb_channels in
PixFmtInfo.

See thread:
Subject: [FFmpeg-devel] [PATCH 6/8] Make avcodec_pix_fmt_string() use the
        information in the pixel format descriptors for printing the
        number of planes. Also replace the term "nb_channels" with
        "nb_planes" which is more correct.
Date: Fri,  5 Nov 2010 12:01:38 +0100

Originally committed as revision 25717 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoIntel C compiler 12.0 does not suport these attributes: may_alias, force_align_arg_po...
Carl Eugen Hoyos [Wed, 10 Nov 2010 12:45:41 +0000 (12:45 +0000)]
Intel C compiler 12.0 does not suport these attributes: may_alias, force_align_arg_pointer and alloc_size.

Originally committed as revision 25716 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMove code to be used by the IEC 61937 demuxer from spifenc.c into common
Anssi Hannula [Wed, 10 Nov 2010 00:04:12 +0000 (00:04 +0000)]
Move code to be used by the IEC 61937 demuxer from spifenc.c into common
files spdif.h and spdif.c.

Patch by Anssi Hannula, anssi d hannula a iki d fi

Originally committed as revision 25715 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoSimplify: Remove impossible condition.
Nicolas Kaiser [Tue, 9 Nov 2010 23:58:41 +0000 (23:58 +0000)]
Simplify: Remove impossible condition.

Patch by Nicolas Kaiser, nikai nikai net

Originally committed as revision 25714 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRename spdif.c as spdifenc,c to prepare for future spdif demuxer.
Anssi Hannula [Tue, 9 Nov 2010 23:57:02 +0000 (23:57 +0000)]
Rename spdif.c as spdifenc,c to prepare for future spdif demuxer.

Patch by Anssi Hannula, anssi d hannula a iki d fi

Originally committed as revision 25713 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMove internal function ff_set_systematic_pal() to libavcore, and
Stefano Sabatini [Tue, 9 Nov 2010 22:22:36 +0000 (22:22 +0000)]
Move internal function ff_set_systematic_pal() to libavcore, and
rename it ff_set_systematic_pal2().

Originally committed as revision 25712 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoReindent after last commit.
Thilo Borgmann [Mon, 8 Nov 2010 15:24:11 +0000 (15:24 +0000)]
Reindent after last commit.

Originally committed as revision 25711 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAllow clean initialization of lut_status array by changing type to signed int.
Thilo Borgmann [Mon, 8 Nov 2010 15:22:15 +0000 (15:22 +0000)]
Allow clean initialization of lut_status array by changing type to signed int.

Originally committed as revision 25710 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd a file for bitstream filters documentation.
Stefano Sabatini [Mon, 8 Nov 2010 14:27:16 +0000 (14:27 +0000)]
Add a file for bitstream filters documentation.

Based on a patch by Nicolas George
<@var{name}.@var{surname}@@normalesup.@code{org}>.

Originally committed as revision 25709 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoUse hierarchic names convention (prefix them with av_expr) for the
Stefano Sabatini [Mon, 8 Nov 2010 14:06:49 +0000 (14:06 +0000)]
Use hierarchic names convention (prefix them with av_expr) for the
eval API.

More grep-friendly and more consistent with the rest of the FFmpeg
API.

Originally committed as revision 25708 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRename avfilter_destroy() as avfilter_free().
Stefano Sabatini [Mon, 8 Nov 2010 12:32:39 +0000 (12:32 +0000)]
Rename avfilter_destroy() as avfilter_free().

The new name is shorter and more consistent with the FFmpeg API, and
sounds less evil.

Originally committed as revision 25707 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoffmpeg.c: allow using negative file indices to disable metadata copying
Anton Khirnov [Mon, 8 Nov 2010 06:47:41 +0000 (06:47 +0000)]
ffmpeg.c: allow using negative file indices to disable metadata copying

Originally committed as revision 25706 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRemove graphparser.h header, move AVFilterInOut and
Stefano Sabatini [Sun, 7 Nov 2010 22:56:29 +0000 (22:56 +0000)]
Remove graphparser.h header, move AVFilterInOut and
avfilter_graph_parse() declaration to libavfilter/avfiltergraph.h.

Simplify, and less confusing for the user.

Originally committed as revision 25705 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoDocument AVFilterInOut fields.
Stefano Sabatini [Sun, 7 Nov 2010 22:47:37 +0000 (22:47 +0000)]
Document AVFilterInOut fields.

Originally committed as revision 25704 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoExtend documentation for parse_filter().
Stefano Sabatini [Sun, 7 Nov 2010 22:47:33 +0000 (22:47 +0000)]
Extend documentation for parse_filter().

Originally committed as revision 25703 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix date in latest APIchanges entry.
Stefano Sabatini [Sun, 7 Nov 2010 19:22:55 +0000 (19:22 +0000)]
Fix date in latest APIchanges entry.

Originally committed as revision 25702 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoBump lavfi minor and add APIchanges entry after the rename of
Stefano Sabatini [Sun, 7 Nov 2010 18:56:44 +0000 (18:56 +0000)]
Bump lavfi minor and add APIchanges entry after the rename of
AVFilterInOut.filter.

Originally committed as revision 25701 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRename AVFilterInOut field filter to filter_ctx, more semantically correct.
Stefano Sabatini [Sun, 7 Nov 2010 18:40:18 +0000 (18:40 +0000)]
Rename AVFilterInOut field filter to filter_ctx, more semantically correct.

Originally committed as revision 25700 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake avfilter_graph_parse() always return meaningful error codes.
Stefano Sabatini [Sun, 7 Nov 2010 18:40:14 +0000 (18:40 +0000)]
Make avfilter_graph_parse() always return meaningful error codes.

Originally committed as revision 25699 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake parse_outputs() always return meaningful error codes.
Stefano Sabatini [Sun, 7 Nov 2010 18:40:12 +0000 (18:40 +0000)]
Make parse_outputs() always return meaningful error codes.

Originally committed as revision 25698 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd missing memory check in parse_inputs(), and fix possible (yet very unlikely)...
Stefano Sabatini [Sun, 7 Nov 2010 18:40:09 +0000 (18:40 +0000)]
Add missing memory check in parse_inputs(), and fix possible (yet very unlikely) crash.

Originally committed as revision 25697 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake parse_inputs() return AVERROR(EINVAL) rather than -1 in case the
Stefano Sabatini [Sun, 7 Nov 2010 18:40:07 +0000 (18:40 +0000)]
Make parse_inputs() return AVERROR(EINVAL) rather than -1 in case the
link label parsed by parse_link_name() is invalid.

Originally committed as revision 25696 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd missing check in link_filter_inouts(), and fix possible (yet very unlikely) crash.
Stefano Sabatini [Sun, 7 Nov 2010 18:40:04 +0000 (18:40 +0000)]
Add missing check in link_filter_inouts(), and fix possible (yet very unlikely) crash.

Originally committed as revision 25695 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRename the variable filter to filt_ctx in link_filter_inouts, as the variable denotes...
Stefano Sabatini [Sun, 7 Nov 2010 18:40:01 +0000 (18:40 +0000)]
Rename the variable filter to filt_ctx in link_filter_inouts, as the variable denotes a filter context.

Less confusing.

Originally committed as revision 25694 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake link_filter_inouts() return meaningful error codes.
Stefano Sabatini [Sun, 7 Nov 2010 18:39:58 +0000 (18:39 +0000)]
Make link_filter_inouts() return meaningful error codes.

Originally committed as revision 25693 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoChange the signature of parse_filter() to make it return an error code.
Stefano Sabatini [Sun, 7 Nov 2010 18:39:54 +0000 (18:39 +0000)]
Change the signature of parse_filter() to make it return an error code.

Originally committed as revision 25692 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoChange the signature of create_filter() to make it return an error code.
Stefano Sabatini [Sun, 7 Nov 2010 18:39:52 +0000 (18:39 +0000)]
Change the signature of create_filter() to make it return an error code.

Originally committed as revision 25691 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake x11grab cursor drawing suck less
Jason Garrett-Glaser [Sun, 7 Nov 2010 18:04:46 +0000 (18:04 +0000)]
Make x11grab cursor drawing suck less
This new version:
1.  Works on 24-bit and 32-bit input, not just 32-bit.
2.  Doesn't try to run on 16-bit or 8-bit, instead of outright crashing.
3.  Does proper alpha-blending, so cursor shadows look correct.
4.  Doesn't swap R and B.

Mostly fixes issue 1997.
Fixes issue 2056.

Originally committed as revision 25690 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoOptimize / simplify avcodec_get_pix_fmt_name() check.
Stefano Sabatini [Sun, 7 Nov 2010 15:04:35 +0000 (15:04 +0000)]
Optimize / simplify avcodec_get_pix_fmt_name() check.

Originally committed as revision 25689 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoUse av_pix_fmt_descriptors information in avcodec_pix_fmt_string().
Stefano Sabatini [Sun, 7 Nov 2010 13:30:39 +0000 (13:30 +0000)]
Use av_pix_fmt_descriptors information in avcodec_pix_fmt_string().

Originally committed as revision 25688 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake avpicture_fill() return a meaningful error code.
Stefano Sabatini [Sun, 7 Nov 2010 13:30:36 +0000 (13:30 +0000)]
Make avpicture_fill() return a meaningful error code.

Originally committed as revision 25687 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRemove alpha information from avcodec_pix_fmt_string(), as that
Stefano Sabatini [Sun, 7 Nov 2010 13:01:12 +0000 (13:01 +0000)]
Remove alpha information from avcodec_pix_fmt_string(), as that
information does not belong to the pixel format.

Originally committed as revision 25686 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaaclatm: Eliminate dummy packets due to muxlength calculation.
Alex Converse [Sun, 7 Nov 2010 03:05:12 +0000 (03:05 +0000)]
aaclatm: Eliminate dummy packets due to muxlength calculation.

Muxlength does not include the 3 bytes of AudioSyncStream() before the
AudioMuxElement(). If these three bytes are not accounted for the last three
bytes of the LATM packet are sent back to the decoder again.

Fixes issue244/mux2.share.ts

Originally committed as revision 25685 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agortsp: Set the real_setup pointer properly, avoid out of bounds memory accesses
Yuriy Kaminskiy [Sat, 6 Nov 2010 23:26:49 +0000 (23:26 +0000)]
rtsp: Set the real_setup pointer properly, avoid out of bounds memory accesses

This fixes a regression since svn rev 24752, where the real_setup pointer
was set incorrectly. The arithmetic with the real_setup_cache pointer
is in units of enum AVDiscard, so the sizeof multiplication should be removed.

Patch by Yuriy Kaminskiy, yumkam at mail dot ru.

Originally committed as revision 25684 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix indentation.
Reimar Döffinger [Sat, 6 Nov 2010 16:48:41 +0000 (16:48 +0000)]
Fix indentation.

Originally committed as revision 25683 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoImprove PGS parsing: the "state" field is not relevant to us,
Reimar Döffinger [Sat, 6 Nov 2010 16:46:03 +0000 (16:46 +0000)]
Improve PGS parsing: the "state" field is not relevant to us,
the object number is, it determines whether we should continue
parsing the presentation description and whether we should
clear the subtitles on the next display command.
Based on patch by Mark Goodman [mark goodman gmail com]

Originally committed as revision 25682 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agotry to give TrueSpeech decoder tables more sensible names
Kostya Shishkov [Sat, 6 Nov 2010 13:31:17 +0000 (13:31 +0000)]
try to give TrueSpeech decoder tables more sensible names

Originally committed as revision 25681 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoadd my (compn) fingerprint to svn write list in maintainers file
Ben Littler [Fri, 5 Nov 2010 16:37:41 +0000 (16:37 +0000)]
add my (compn) fingerprint to svn write list in maintainers file

Originally committed as revision 25680 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoImplement get_preset_file() in cmdutils.h and use it to factorize code
Stefano Sabatini [Thu, 4 Nov 2010 23:07:04 +0000 (23:07 +0000)]
Implement get_preset_file() in cmdutils.h and use it to factorize code
from ffmpeg.c and ffserver.c.

Originally committed as revision 25679 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix FFv1 decoder buffer releasing.
Stefano Sabatini [Thu, 4 Nov 2010 22:45:46 +0000 (22:45 +0000)]
Fix FFv1 decoder buffer releasing.

Previously it was releasing the buffer which was returned to the user,
which was resulting in a crash in case of direct rendering.

Originally committed as revision 25678 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd test for cropping of interlaced H.264.
Reimar Döffinger [Thu, 4 Nov 2010 21:48:54 +0000 (21:48 +0000)]
Add test for cropping of interlaced H.264.

Originally committed as revision 25677 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: consume the audio specific config during LATM parsing
Janne Grunau [Thu, 4 Nov 2010 21:00:01 +0000 (21:00 +0000)]
aacdec: consume the audio specific config during LATM parsing

Spotted by Alex after Carl Eugen found errors some samples. There no errors or
noticeable artifacts in the samples I used during development.

Originally committed as revision 25676 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix off-by-one error in APIchanges file.
Stefano Sabatini [Thu, 4 Nov 2010 20:37:52 +0000 (20:37 +0000)]
Fix off-by-one error in APIchanges file.

Originally committed as revision 25675 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRename avfilter_graph_destroy() to avfilter_graph_free().
Stefano Sabatini [Thu, 4 Nov 2010 20:34:24 +0000 (20:34 +0000)]
Rename avfilter_graph_destroy() to avfilter_graph_free().

The new name is shorter and more consistent with the rest of the API.

This change breaks libavfilter API/ABI.

Originally committed as revision 25674 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoImplement avfilter_graph_alloc().
Stefano Sabatini [Thu, 4 Nov 2010 20:34:19 +0000 (20:34 +0000)]
Implement avfilter_graph_alloc().

Originally committed as revision 25673 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAmend avfilter_graph_parse() documentation for the return code, make
Stefano Sabatini [Thu, 4 Nov 2010 20:26:45 +0000 (20:26 +0000)]
Amend avfilter_graph_parse() documentation for the return code, make
it more generic.

Originally committed as revision 25672 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRemove two instances of AVERROR(ff_neterrno()), because ff_neterrno() itself
Ronald S. Bultje [Thu, 4 Nov 2010 13:53:16 +0000 (13:53 +0000)]
Remove two instances of AVERROR(ff_neterrno()), because ff_neterrno() itself
already does AVERROR().

Originally committed as revision 25671 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoSilence icc warning about unsupported option -Wno-parentheses
Carl Eugen Hoyos [Thu, 4 Nov 2010 10:38:39 +0000 (10:38 +0000)]
Silence icc warning about unsupported option -Wno-parentheses

Originally committed as revision 25670 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoH.264 decode: support cropping up to 28 pixels in interlaced mode.
Reimar Döffinger [Wed, 3 Nov 2010 20:52:28 +0000 (20:52 +0000)]
H.264 decode: support cropping up to 28 pixels in interlaced mode.
Contrary to progressive, just being able to crop up to 14/15 pixels
is not enough to encode all supported resolutions, and the new
behaviour is also consistent with e.g. MPEG-2 etc.

Originally committed as revision 25669 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRemove usage of deprecated libavcodec/audioconvert.h functions.
Stefano Sabatini [Wed, 3 Nov 2010 20:19:34 +0000 (20:19 +0000)]
Remove usage of deprecated libavcodec/audioconvert.h functions.

Originally committed as revision 25668 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix 16kHz Mono NellyMoser in flv.
Thierry Foucu [Wed, 3 Nov 2010 20:12:07 +0000 (20:12 +0000)]
Fix 16kHz Mono NellyMoser in flv.

Patch by Thierry Foucu, tfoucu gmail

Originally committed as revision 25667 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoImplement isnan() function evaluation.
Stefano Sabatini [Wed, 3 Nov 2010 19:44:00 +0000 (19:44 +0000)]
Implement isnan() function evaluation.

Originally committed as revision 25666 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd eval.texi file.
Stefano Sabatini [Wed, 3 Nov 2010 19:43:55 +0000 (19:43 +0000)]
Add eval.texi file.

Originally committed as revision 25665 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoUpdate 24 bpp TM1 reference for decoder fixes.
Reimar Döffinger [Wed, 3 Nov 2010 18:13:09 +0000 (18:13 +0000)]
Update 24 bpp TM1 reference for decoder fixes.

Originally committed as revision 25664 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix aspect for 24bpp TM1 samples.
Reimar Döffinger [Wed, 3 Nov 2010 17:22:19 +0000 (17:22 +0000)]
Fix aspect for 24bpp TM1 samples.

Originally committed as revision 25663 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix decoding of 24 bpp TM1 (except for aspect).
Reimar Döffinger [Wed, 3 Nov 2010 17:01:30 +0000 (17:01 +0000)]
Fix decoding of 24 bpp TM1 (except for aspect).

Originally committed as revision 25662 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years ago100l, calculate mb_change_bits_row_size from the update width.
Reimar Döffinger [Wed, 3 Nov 2010 16:57:05 +0000 (16:57 +0000)]
100l, calculate mb_change_bits_row_size from the update width.

Originally committed as revision 25661 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoSupport mid-stream resolution/format changes for TM1.
Reimar Döffinger [Wed, 3 Nov 2010 16:45:55 +0000 (16:45 +0000)]
Support mid-stream resolution/format changes for TM1.
This makes it easier to fix playback of 24 bit formats.

Originally committed as revision 25660 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agomatroskadec: skip invalid tags with no TagName
Anton Khirnov [Wed, 3 Nov 2010 06:29:04 +0000 (06:29 +0000)]
matroskadec: skip invalid tags with no TagName

fixes segfault in issue 2328

Originally committed as revision 25659 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agochangelog: mention that ffmpeg is now win64-friendly
Ramiro Polla [Tue, 2 Nov 2010 23:26:23 +0000 (23:26 +0000)]
changelog: mention that ffmpeg is now win64-friendly

Originally committed as revision 25658 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years ago1/PHIl: swap again settb/setpts, fix alphabetical order.
Stefano Sabatini [Tue, 2 Nov 2010 22:53:18 +0000 (22:53 +0000)]
1/PHIl: swap again settb/setpts, fix alphabetical order.

Originally committed as revision 25657 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMove settb docs after setpts docs, fix alphabetical order.
Stefano Sabatini [Tue, 2 Nov 2010 22:40:31 +0000 (22:40 +0000)]
Move settb docs after setpts docs, fix alphabetical order.

Originally committed as revision 25656 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd setpts filter, ported from the libavfilter soc repo.
Stefano Sabatini [Tue, 2 Nov 2010 22:29:17 +0000 (22:29 +0000)]
Add setpts filter, ported from the libavfilter soc repo.

Originally committed as revision 25655 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd av_get_bits_per_sample_fmt() to libavcore/samplefmt.h and
Stefano Sabatini [Tue, 2 Nov 2010 22:20:49 +0000 (22:20 +0000)]
Add av_get_bits_per_sample_fmt() to libavcore/samplefmt.h and
deprecate av_get_bits_per_sample_format().

Originally committed as revision 25654 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoImplement functions:
Stefano Sabatini [Tue, 2 Nov 2010 22:08:02 +0000 (22:08 +0000)]
Implement functions:
av_get_sample_fmt_name()
av_get_sample_fmt()
av_get_sample_fmt_string()

in libavcore, and deprecate the corresponding libavcodec/audioconvert.h functions:
avcodec_get_sample_fmt_name()
avcodec_get_sample_fmt()
avcodec_sample_fmt_string()

Originally committed as revision 25653 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMove sample format definitions from libavcodec to libavcore.
Stefano Sabatini [Tue, 2 Nov 2010 20:45:46 +0000 (20:45 +0000)]
Move sample format definitions from libavcodec to libavcore.

Originally committed as revision 25652 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agohttp: Handle 301/307 redirect requests, too
Mike Edenfield [Tue, 2 Nov 2010 20:37:01 +0000 (20:37 +0000)]
http: Handle 301/307 redirect requests, too

Patch by Mike Edenfield, kutulu at kutulu dot org

Originally committed as revision 25651 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAllow mono encoding with LAME.
Thierry Foucu [Tue, 2 Nov 2010 20:29:20 +0000 (20:29 +0000)]
Allow mono encoding with LAME.

Patch by Thierry Foucu, tfoucu gmail

Originally committed as revision 25650 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix timestamp handling in more filters after the change which set the
Stefano Sabatini [Tue, 2 Nov 2010 19:42:08 +0000 (19:42 +0000)]
Fix timestamp handling in more filters after the change which set the
timebase in the filter link.

Originally committed as revision 25649 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake the blackframe filter print the time, in addition to the
Stefano Sabatini [Tue, 2 Nov 2010 19:42:05 +0000 (19:42 +0000)]
Make the blackframe filter print the time, in addition to the
timestamp, for each detected black frame.

Originally committed as revision 25648 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake the cropdetect filter print the time for each frame, in addition
Stefano Sabatini [Tue, 2 Nov 2010 19:42:03 +0000 (19:42 +0000)]
Make the cropdetect filter print the time for each frame, in addition
to the timestamp.

Originally committed as revision 25647 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix timestamp computation when the timebase changes during filtering.
Stefano Sabatini [Tue, 2 Nov 2010 18:40:59 +0000 (18:40 +0000)]
Fix timestamp computation when the timebase changes during filtering.

The timestamps for all the buffered frames are scaled against
AV_TIME_BASE_Q, and need to be scaled back to AV_TIME_BASE_Q when they
are extracted from the filterchain.

Originally committed as revision 25646 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix crop->var_values[VAR_T] computation in the crop filter.
Stefano Sabatini [Tue, 2 Nov 2010 18:40:55 +0000 (18:40 +0000)]
Fix crop->var_values[VAR_T] computation in the crop filter.

Originally committed as revision 25645 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: change type of data in decode_audio_specific_config parameters
Janne Grunau [Tue, 2 Nov 2010 08:33:10 +0000 (08:33 +0000)]
aacdec: change type of data in decode_audio_specific_config parameters

AVCodecContext.extradata is uint8_t*, silence a warning

Originally committed as revision 25644 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agompegts: support LATM syntax
Janne Grunau [Tue, 2 Nov 2010 08:32:33 +0000 (08:32 +0000)]
mpegts: support LATM syntax

Originally committed as revision 25643 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd single stream LATM/LOAS decoder
Janne Grunau [Tue, 2 Nov 2010 08:32:04 +0000 (08:32 +0000)]
Add single stream LATM/LOAS decoder

The decoder is just a wrapper around the AAC decoder.
based on patch by Paul Kendall { paul <ät> kcbbs gen nz }

Originally committed as revision 25642 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: pass avctx as logging context for decode_audio_specific_config
Janne Grunau [Tue, 2 Nov 2010 08:30:31 +0000 (08:30 +0000)]
aacdec: pass avctx as logging context for decode_audio_specific_config

Use avctx in all called functions. This allows passing a NULL AACContext
for LATM since the AACContext is only used in output_configure() which
is skipped for LATM parsing.

Originally committed as revision 25641 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: refactor the actual aac decoding code into its own function
Janne Grunau [Tue, 2 Nov 2010 08:29:43 +0000 (08:29 +0000)]
aacdec: refactor the actual aac decoding code into its own function

aac_decode_frame() remains as AVPacket handling a wrapper. The actual
decoding function takes a GetBitContext as input and will be used be the
AAC LATM decoder to avoid copying the unaligned AAC bitstream.

Originally committed as revision 25640 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: return consumed bits in decode_audio_specific_config
Janne Grunau [Tue, 2 Nov 2010 08:28:55 +0000 (08:28 +0000)]
aacdec: return consumed bits in decode_audio_specific_config

Originally committed as revision 25639 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacdec: add MPEG4AudioConfig as parameter for decode_audio_specific_config
Janne Grunau [Tue, 2 Nov 2010 08:28:28 +0000 (08:28 +0000)]
aacdec: add MPEG4AudioConfig as parameter for decode_audio_specific_config

This will be used by the latm decoder to avoid AACContext changes during
audio specific config parsing.

Originally committed as revision 25638 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoffmpeg.c: reindent
Anton Khirnov [Tue, 2 Nov 2010 06:12:56 +0000 (06:12 +0000)]
ffmpeg.c: reindent

Originally committed as revision 25637 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoffmpeg.c: don't autocopy stream/chapter metadata if manual mapping is specified
Anton Khirnov [Tue, 2 Nov 2010 06:12:54 +0000 (06:12 +0000)]
ffmpeg.c: don't autocopy stream/chapter metadata if manual mapping is specified

Originally committed as revision 25636 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoffmpeg.c: extend map_meta_data to allow advanced mappings
Anton Khirnov [Tue, 2 Nov 2010 06:01:28 +0000 (06:01 +0000)]
ffmpeg.c: extend map_meta_data to allow advanced mappings

i.e. to/from streams/chapters/programs.

Originally committed as revision 25635 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agodnxhd_mmx: prefer xmm registers below xmm6 when they are available
Ramiro Polla [Tue, 2 Nov 2010 03:09:16 +0000 (03:09 +0000)]
dnxhd_mmx: prefer xmm registers below xmm6 when they are available

Originally committed as revision 25634 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix possibly exploitable out of buffer writes in msrle_decode_pal4().
Michael Niedermayer [Tue, 2 Nov 2010 01:19:15 +0000 (01:19 +0000)]
Fix possibly exploitable out of buffer writes in msrle_decode_pal4().
This fix is minimalistic, that function should be cleaned up by someone.

Originally committed as revision 25633 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix possibly exploitable buffer overrun in msrle_decode_8_16_24_32().
Michael Niedermayer [Tue, 2 Nov 2010 01:19:12 +0000 (01:19 +0000)]
Fix possibly exploitable buffer overrun in msrle_decode_8_16_24_32().
Issue has been reported to me by Gynvael Coldwind

Originally committed as revision 25632 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoremove impossible condition from msrle_decode_pal4()
Michael Niedermayer [Tue, 2 Nov 2010 01:19:09 +0000 (01:19 +0000)]
remove impossible condition from msrle_decode_pal4()

Originally committed as revision 25631 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agofix indent
Michael Niedermayer [Tue, 2 Nov 2010 01:19:03 +0000 (01:19 +0000)]
fix indent

Originally committed as revision 25630 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoRemove ffmpeg PGMYUV compatibility hack, which was deprecated since
Stefano Sabatini [Tue, 2 Nov 2010 00:29:13 +0000 (00:29 +0000)]
Remove ffmpeg PGMYUV compatibility hack, which was deprecated since
ages.

The user is requested to specify "-f image2" in place of "-f pgmyuv"
for reading/writing PGMYUV files, as for the other image formats.

Originally committed as revision 25629 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agodsputil: Use explicit movzbl instead of movzx
İsmail Dönmez [Mon, 1 Nov 2010 19:35:51 +0000 (19:35 +0000)]
dsputil: Use explicit movzbl instead of movzx

This fixes compilation with the latest clang trunk version.

Patch by İsmail Dönmez, ismail at namtrac dot org

Originally committed as revision 25628 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix documented syntax for the cropdetect filter.
Stefano Sabatini [Mon, 1 Nov 2010 10:33:10 +0000 (10:33 +0000)]
Fix documented syntax for the cropdetect filter.

Originally committed as revision 25627 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoMake strmatch() return 1 only if the string compared against the
Stefano Sabatini [Mon, 1 Nov 2010 09:34:21 +0000 (09:34 +0000)]
Make strmatch() return 1 only if the string compared against the
prefix does not contain other characters which may belong to an
identifier.

This allows to distinguish for example to have different constants
with the same prefix (e.g. "foo" and "foobar").

Originally committed as revision 25626 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd st, ld, while examples/tests.
Stefano Sabatini [Mon, 1 Nov 2010 09:34:18 +0000 (09:34 +0000)]
Add st, ld, while examples/tests.

Originally committed as revision 25625 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoAdd examples of unary operators.
Stefano Sabatini [Mon, 1 Nov 2010 09:34:15 +0000 (09:34 +0000)]
Add examples of unary operators.

Originally committed as revision 25624 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoFix evaluation of expressions of the form: E1;E2.
Stefano Sabatini [Mon, 1 Nov 2010 09:34:12 +0000 (09:34 +0000)]
Fix evaluation of expressions of the form: E1;E2.

The pointer to the char ';' has to be increased before to evaluate
";E2".

Originally committed as revision 25623 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacenc: Correct spreading calculation for high spreading.
Nathan Caldwell [Mon, 1 Nov 2010 07:45:13 +0000 (07:45 +0000)]
aacenc: Correct spreading calculation for high spreading.

The 3GPP spec uses the following calculation for high spreading:

thr'_spr = max(thr_scaled, s_h(n) * thr_scaled(n-1))

where, n is defined as the current band, and s_h() is defined as "[...] the
distance of adjacent bands in Bark and a constant slope that is 15 dB/Bark
[...]". This is a little ambiguous as you would assume you want the Bark
width of the previous band for this calculation. However, this assumption
appears to be incorrect, and you really want the Bark width of the current
band. Coincidentally this is exactly what the spec calls for! =P

This noticeably improves Tom's Diner at low bitrates (I tested at 64kbps,
with mid/side disabled).

Patch by: Nathan Caldwell <saintdev@gmail.com>

Originally committed as revision 25622 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agoaacenc: cosmetics: Swap spreading_hi/low name to match the 3GPP spec.
Nathan Caldwell [Mon, 1 Nov 2010 07:39:31 +0000 (07:39 +0000)]
aacenc: cosmetics: Swap spreading_hi/low name to match the 3GPP spec.

Patch by: Nathan Caldwell <saintdev@gmail.com>

Originally committed as revision 25621 to svn://svn.ffmpeg.org/ffmpeg/trunk

8 years agolpc_mmx: add xmm registers to clobber list
Ramiro Polla [Sun, 31 Oct 2010 23:37:15 +0000 (23:37 +0000)]
lpc_mmx: add xmm registers to clobber list

Originally committed as revision 25620 to svn://svn.ffmpeg.org/ffmpeg/trunk