12 years agoReplace wrong condition name in #endif comment by correct instance.
Diego Biurrun [Fri, 19 Mar 2010 09:58:45 +0000 (09:58 +0000)]
Replace wrong condition name in #endif comment by correct instance.

Originally committed as revision 30938 to svn://

12 years agoCheck whether COMPILE_ALTIVEC is defined, not if it is set to a 0/1 value.
Diego Biurrun [Fri, 19 Mar 2010 09:57:36 +0000 (09:57 +0000)]
Check whether COMPILE_ALTIVEC is defined, not if it is set to a 0/1 value.
COMPILE_ALTIVEC is never set to 1, it is just #defined.

Originally committed as revision 30937 to svn://

12 years agoffmpeg.c: copy chapters by default.
Anton Khirnov [Fri, 19 Mar 2010 07:44:38 +0000 (07:44 +0000)]
ffmpeg.c: copy chapters by default.
Patch by Anton Khirnov wyskas gmail com

Originally committed as revision 22598 to svn://

12 years agoReinitialize the h264 decoder context on every valid aspect ratio
Janusz Krzysztofik [Thu, 18 Mar 2010 23:36:54 +0000 (23:36 +0000)]
Reinitialize the h264 decoder context on every valid aspect ratio
change, not only size changes.

Patch by Janusz Krzysztofik foo=zyszt <jkr$>.

Originally committed as revision 22597 to svn://

12 years agoThe vflip filter does not need to keep ownership of a picture reference.
Bobby Bingham [Thu, 18 Mar 2010 23:13:04 +0000 (23:13 +0000)]
The vflip filter does not need to keep ownership of a picture reference.
Avoid creating and releasing useless extra picture references in the
start_frame and end_frame callbacks.

Originally committed as revision 22596 to svn://

12 years agoUse the new null callbacks to simplify filters and reduce duplicated code.
Bobby Bingham [Thu, 18 Mar 2010 23:13:02 +0000 (23:13 +0000)]
Use the new null callbacks to simplify filters and reduce duplicated code.

Originally committed as revision 22595 to svn://

12 years agoCreate a set of null callback functions.
Bobby Bingham [Thu, 18 Mar 2010 23:12:48 +0000 (23:12 +0000)]
Create a set of null callback functions.
These are useful for filters which don't modify the image data.

Originally committed as revision 22594 to svn://

12 years agoExtend the generic path of the yuv2rgb converter with support for rgb444
Janusz Krzysztofik [Thu, 18 Mar 2010 23:02:32 +0000 (23:02 +0000)]
Extend the generic path of the yuv2rgb converter with support for rgb444
output format.

Patch by Janusz Krzysztofik, jkrzyszt A tis D icnet D pl

Originally committed as revision 30934 to svn://

12 years agoReplace remaining uses of ATTR_ALIGNED with DECLARE_ALIGNED
Måns Rullgård [Thu, 18 Mar 2010 15:00:17 +0000 (15:00 +0000)]
Replace remaining uses of ATTR_ALIGNED with DECLARE_ALIGNED

Originally committed as revision 22593 to svn://

12 years agoIt is difficult nowadays to compile without swscale support.
Carl Eugen Hoyos [Thu, 18 Mar 2010 00:24:26 +0000 (00:24 +0000)]
It is difficult nowadays to compile without swscale support.

Originally committed as revision 22592 to svn://

12 years agoMake g729dec.c:decoder_init() return AVERROR(EINVAL) rather than
Stefano Sabatini [Thu, 18 Mar 2010 00:23:02 +0000 (00:23 +0000)]
Make g729dec.c:decoder_init() return AVERROR(EINVAL) rather than
AVERROR(NOFMT) if one channel if the number of channels specified
is different than 1.

Originally committed as revision 22591 to svn://

12 years agoMake url_fseek() return AVERROR_EOF rather than AVERROR(EPIPE) if end
Stefano Sabatini [Thu, 18 Mar 2010 00:22:58 +0000 (00:22 +0000)]
Make url_fseek() return AVERROR_EOF rather than AVERROR(EPIPE) if end
of file is reached.

Originally committed as revision 22590 to svn://

12 years agoRemove printing of frame sizes and frame rate abbreviations from
Stefano Sabatini [Wed, 17 Mar 2010 23:51:12 +0000 (23:51 +0000)]
Remove printing of frame sizes and frame rate abbreviations from

The list of abbreviations is both outdated and out of context.

Originally committed as revision 22589 to svn://

12 years agoAdd a -window_title option, which sets the FFplay window title.
Robert Krüger [Wed, 17 Mar 2010 23:39:18 +0000 (23:39 +0000)]
Add a -window_title option, which sets the FFplay window title.

Patch by Robert Krüger "krueger ET signal7 DOT de".

Originally committed as revision 22588 to svn://

12 years agoMake ad-hoc lavfi tests use random values for the slice height used
Stefano Sabatini [Wed, 17 Mar 2010 23:23:48 +0000 (23:23 +0000)]
Make ad-hoc lavfi tests use random values for the slice height used
per each frame, useful for testing slicification.

Originally committed as revision 22587 to svn://

12 years agoAdd some ad-hoc tests for libavfilter.
Stefano Sabatini [Wed, 17 Mar 2010 23:23:44 +0000 (23:23 +0000)]
Add some ad-hoc tests for libavfilter.

A patched version of ffmpeg supporting video filters is required for
getting this working; thus make lavfitest is supposed to work only in
the libavfilter repository for now.

Originally committed as revision 22586 to svn://

12 years agoCast sample_fmt (as read from bitstream as a 16-bit value) to int16, so that
Ronald S. Bultje [Wed, 17 Mar 2010 15:14:18 +0000 (15:14 +0000)]
Cast sample_fmt (as read from bitstream as a 16-bit value) to int16, so that
negative values, e.g. SAMPLE_FMT_NONE (-1), are read correctly also.

Originally committed as revision 22585 to svn://

12 years agoFix compilation on PPC 470.
Carl Eugen Hoyos [Wed, 17 Mar 2010 15:06:27 +0000 (15:06 +0000)]
Fix compilation on PPC 470.

Originally committed as revision 22584 to svn://

12 years agoRemove $(MSG) override for install-lib targets
Måns Rullgård [Wed, 17 Mar 2010 14:43:05 +0000 (14:43 +0000)]
Remove $(MSG) override for install-lib targets

This is no longer needed and causes incorrect output in some cases.

Originally committed as revision 22583 to svn://

12 years agoMake ranlib silent in brief output mode
Måns Rullgård [Wed, 17 Mar 2010 14:43:03 +0000 (14:43 +0000)]
Make ranlib silent in brief output mode

Running ranlib is quick and uninteresting, no need to print it.

Originally committed as revision 22582 to svn://

12 years agoqtrleenc: add rgb555 output
Alexis Ballier [Wed, 17 Mar 2010 13:04:33 +0000 (13:04 +0000)]
qtrleenc: add rgb555 output
Patch by Alexis Ballier std_gmail_addr(name.surname)

Originally committed as revision 22581 to svn://

12 years agoCorrect reference buffer switching in Indeo 5 decoder.
Maxim Poliakovski [Wed, 17 Mar 2010 07:53:12 +0000 (07:53 +0000)]
Correct reference buffer switching in Indeo 5 decoder.
Patch by Maxim ($indeo5decauthor)

Originally committed as revision 22580 to svn://

12 years agodocument two variables in Indeo5 decoding context
Kostya Shishkov [Wed, 17 Mar 2010 07:49:36 +0000 (07:49 +0000)]
document two variables in Indeo5 decoding context

Originally committed as revision 22579 to svn://

12 years agoCosmetics: vertical align
Bobby Bingham [Wed, 17 Mar 2010 04:37:43 +0000 (04:37 +0000)]
Cosmetics: vertical align

Originally committed as revision 22578 to svn://

12 years agoAllow format and noformat filters to be more completely disabled when the
Bobby Bingham [Wed, 17 Mar 2010 04:37:40 +0000 (04:37 +0000)]
Allow format and noformat filters to be more completely disabled when the
other is enabled.

Originally committed as revision 22577 to svn://

12 years agoCosmetics: rearrange functions to simplify ifdeffery
Bobby Bingham [Wed, 17 Mar 2010 04:37:36 +0000 (04:37 +0000)]
Cosmetics: rearrange functions to simplify ifdeffery

Originally committed as revision 22576 to svn://

12 years agovfwcap: Free remaining packets on vfw_read_close().
Ramiro Polla [Wed, 17 Mar 2010 04:36:06 +0000 (04:36 +0000)]
vfwcap: Free remaining packets on vfw_read_close().

Originally committed as revision 22575 to svn://

12 years agoDo not store a copy of AVFormatContext *s in vfw_ctx.
Ramiro Polla [Wed, 17 Mar 2010 04:24:43 +0000 (04:24 +0000)]
Do not store a copy of AVFormatContext *s in vfw_ctx.
Save *s directly in SetWindowLongPtr instead.

Originally committed as revision 22574 to svn://

12 years agoAdd video filter to manipulate aspect ratio
Bobby Bingham [Wed, 17 Mar 2010 03:43:14 +0000 (03:43 +0000)]
Add video filter to manipulate aspect ratio

Originally committed as revision 22573 to svn://

12 years agoUse H264s MC instead of mpeg4-asp qpel for h264 error resilience.
Michael Niedermayer [Wed, 17 Mar 2010 02:59:12 +0000 (02:59 +0000)]
Use H264s MC instead of mpeg4-asp qpel for h264 error resilience.

Originally committed as revision 22572 to svn://

12 years agoFix erroneous behaviour when format probe hits end of file
Måns Rullgård [Tue, 16 Mar 2010 21:45:30 +0000 (21:45 +0000)]
Fix erroneous behaviour when format probe hits end of file

If the format probe hits end of file, do not add the error code
to the buffer position.  This is obviously wrong, and with a
small input file would cause a negative buffer overflow.

Fixes issue 1818.

Originally committed as revision 22571 to svn://

12 years agoSimplify arch-specific object file lists
Måns Rullgård [Tue, 16 Mar 2010 21:23:03 +0000 (21:23 +0000)]
Simplify arch-specific object file lists

Originally committed as revision 22570 to svn://

12 years agoMove arch-specific makefile parts into $arch/Makefile
Måns Rullgård [Tue, 16 Mar 2010 21:22:59 +0000 (21:22 +0000)]
Move arch-specific makefile parts into $arch/Makefile

Originally committed as revision 22569 to svn://

12 years agoconfigure: do not use $_
Måns Rullgård [Tue, 16 Mar 2010 21:22:54 +0000 (21:22 +0000)]
configure: do not use $_

Both bash and ksh abuse this variable even in POSIX mode, leading
to unpleasant surprises.

Originally committed as revision 22568 to svn://

12 years agoAdd compilation hints for Cygwin 1.7 and remove the ones for 1.5
Víctor Paesa [Tue, 16 Mar 2010 18:32:52 +0000 (18:32 +0000)]
Add compilation hints for Cygwin 1.7 and remove the ones for 1.5

Originally committed as revision 22567 to svn://

12 years agoTypo: if output (video) stream's pix_fmt is not set, then the stream cannot
Ronald S. Bultje [Tue, 16 Mar 2010 15:48:48 +0000 (15:48 +0000)]
Typo: if output (video) stream's pix_fmt is not set, then the stream cannot
be ENcoded, not DEcoded.

Originally committed as revision 22566 to svn://

12 years agoMove H264 dsputil functions into their own struct
Måns Rullgård [Tue, 16 Mar 2010 01:17:00 +0000 (01:17 +0000)]
Move H264 dsputil functions into their own struct

This moves the H264-specific functions from DSPContext to the new
H264DSPContext.  The code is made conditional on CONFIG_H264DSP
which is set by the codecs requiring it.

The qpel and chroma MC functions are not moved as these are used by
non-h264 code.

Originally committed as revision 22565 to svn://

12 years ago100L, revert r22560, already present
Baptiste Coudurier [Mon, 15 Mar 2010 23:40:51 +0000 (23:40 +0000)]
100L, revert r22560, already present

Originally committed as revision 22564 to svn://

12 years agopatcheck: Escape parentheses in grep calls
Daniel Verkamp [Mon, 15 Mar 2010 23:32:20 +0000 (23:32 +0000)]
patcheck: Escape parentheses in grep calls

Originally committed as revision 22563 to svn://

12 years agorename av_read_frame_flush to ff_read_frame_flush
Aurelien Jacobs [Mon, 15 Mar 2010 23:15:24 +0000 (23:15 +0000)]
rename av_read_frame_flush to ff_read_frame_flush
it is an internal function, not part of public API

Originally committed as revision 22562 to svn://

12 years agorename av_program_add_stream_index to ff_program_add_stream_index
Aurelien Jacobs [Mon, 15 Mar 2010 23:14:07 +0000 (23:14 +0000)]
rename av_program_add_stream_index to ff_program_add_stream_index
it is an internal function, not part of public API

Originally committed as revision 22561 to svn://

12 years agompegts muxer now needs mpeg4audio code like adts muxer
Baptiste Coudurier [Mon, 15 Mar 2010 23:11:02 +0000 (23:11 +0000)]
mpegts muxer now needs mpeg4audio code like adts muxer

Originally committed as revision 22560 to svn://

12 years agoCorrectly write last 0 lacing value when packet size is multiple of 255, patch by...
Greg Maxwell [Mon, 15 Mar 2010 23:04:09 +0000 (23:04 +0000)]
Correctly write last 0 lacing value when packet size is multiple of 255, patch by Greg Maxwell, gmaxwell at gmail dot com

Originally committed as revision 22559 to svn://

12 years agoH264: fix signed overflow in constant multiplication
Måns Rullgård [Mon, 15 Mar 2010 23:00:53 +0000 (23:00 +0000)]
H264: fix signed overflow in constant multiplication

This fixes libavcodec/h264.h:1100: warning: integer overflow in expression

Originally committed as revision 22558 to svn://

12 years agoAdd AVSEEK_FORCE flag to indicate that the code should attempt to seek
Michael Niedermayer [Mon, 15 Mar 2010 22:54:22 +0000 (22:54 +0000)]
Add AVSEEK_FORCE flag to indicate that the code should attempt to seek
by any means.

Originally committed as revision 22557 to svn://

12 years agoAdd ff_ prefix for mpeg2_dc_scale_table.
Michael Niedermayer [Mon, 15 Mar 2010 22:49:46 +0000 (22:49 +0000)]
Add ff_ prefix for mpeg2_dc_scale_table.

Originally committed as revision 22556 to svn://

12 years agomatroskadec: fix missing dependency
Aurelien Jacobs [Mon, 15 Mar 2010 22:37:14 +0000 (22:37 +0000)]
matroskadec: fix missing dependency

Originally committed as revision 22555 to svn://

12 years agoSupport intra_dc_precision>8 in jpeg
Michael Niedermayer [Mon, 15 Mar 2010 21:27:13 +0000 (21:27 +0000)]
Support intra_dc_precision>8 in jpeg

Originally committed as revision 22554 to svn://

12 years agoAllow mpeg style yuv in jpeg when strict standard compliance is small enough.
Michael Niedermayer [Mon, 15 Mar 2010 19:45:57 +0000 (19:45 +0000)]
Allow mpeg style yuv in jpeg when strict standard compliance is small enough.

Originally committed as revision 22553 to svn://

12 years agoAdd FATE tests
Måns Rullgård [Mon, 15 Mar 2010 19:23:24 +0000 (19:23 +0000)]
Add FATE tests

This adds a "fate" make target which runs the full FATE test suite.
Individual tests can be run with "make fate-$testname".

The location of the FATE test samples must be specified with the
--samples=PATH option to configure.

The tests/ script regenerates the references files and
test list from the online FATE database.  These are checked in since
generating them requires non-standard tools.

Originally committed as revision 22552 to svn://

12 years agocosmetics : Print newline after error message.
Jai Menon [Mon, 15 Mar 2010 19:03:48 +0000 (19:03 +0000)]
cosmetics : Print newline after error message.

Originally committed as revision 22551 to svn://

12 years agouse mpeg2 quantization bias for mjpeg.
Michael Niedermayer [Mon, 15 Mar 2010 16:37:02 +0000 (16:37 +0000)]
use mpeg2 quantization bias for mjpeg.
this seems to improve RD performance.

Originally committed as revision 22550 to svn://

12 years agoReindent
Martin Storsjö [Mon, 15 Mar 2010 16:36:54 +0000 (16:36 +0000)]

Originally committed as revision 22549 to svn://

12 years agoDon't let ff_rtsp_read_reply skip interleaved RTP/TCP packets in rtsp_write_packet.
Martin Storsjö [Mon, 15 Mar 2010 16:36:20 +0000 (16:36 +0000)]
Don't let ff_rtsp_read_reply skip interleaved RTP/TCP packets in rtsp_write_packet.
Skip interleaved packets manually and recheck if there's more to be read.

Originally committed as revision 22548 to svn://

12 years agoMake rtsp_skip_packet non-static, add ff prefix
Martin Storsjö [Mon, 15 Mar 2010 16:31:15 +0000 (16:31 +0000)]
Make rtsp_skip_packet non-static, add ff prefix

Originally committed as revision 22547 to svn://

12 years agoReindent
Martin Storsjö [Mon, 15 Mar 2010 16:15:51 +0000 (16:15 +0000)]

Originally committed as revision 22546 to svn://

12 years agoRTSP: Synchronize the start time of the chained RTP muxers
Martin Storsjö [Mon, 15 Mar 2010 14:20:07 +0000 (14:20 +0000)]
RTSP: Synchronize the start time of the chained RTP muxers

This makes sure that the streams get correctly synchronized when viewed,
previously the streams were out of sync by as much time as it took
between the initialization of the individual muxers.

Originally committed as revision 22545 to svn://

12 years agoRemove avcodec_thread_execute from avcodec.h, and make static functions that
Benoit Fouet [Mon, 15 Mar 2010 14:03:01 +0000 (14:03 +0000)]
Remove avcodec_thread_execute from avcodec.h, and make static functions that
need it in *thread.c.

Originally committed as revision 22544 to svn://

12 years agoUse AVFormatContext.start_time_realtime in the RTP muxer
Martin Storsjö [Mon, 15 Mar 2010 10:37:12 +0000 (10:37 +0000)]
Use AVFormatContext.start_time_realtime in the RTP muxer

Originally committed as revision 22543 to svn://

12 years agoMove the NTP offset definitions to internal.h
Martin Storsjö [Mon, 15 Mar 2010 10:36:30 +0000 (10:36 +0000)]
Move the NTP offset definitions to internal.h

Originally committed as revision 22542 to svn://

12 years agoAdd APIchanges entry for AVFormatContext.start_time_realtime
Martin Storsjö [Mon, 15 Mar 2010 10:30:38 +0000 (10:30 +0000)]
Add APIchanges entry for AVFormatContext.start_time_realtime

Originally committed as revision 22541 to svn://

12 years agoAdd a new field AVFormatContext.start_time_realtime
Martin Storsjö [Mon, 15 Mar 2010 10:29:37 +0000 (10:29 +0000)]
Add a new field AVFormatContext.start_time_realtime

Currently intended to be used by the RTP muxer

Originally committed as revision 22540 to svn://

12 years agoriff: don't pad extradata when writing ASF.
Anton Khirnov [Mon, 15 Mar 2010 08:49:02 +0000 (08:49 +0000)]
riff: don't pad extradata when writing ASF.
Patch by Anton Khirnov mirror(moc liamg saksyw)

Originally committed as revision 22539 to svn://

12 years agoAdd dependencies used by the RDT and ASF/RTP code
Martin Storsjö [Mon, 15 Mar 2010 08:35:07 +0000 (08:35 +0000)]
Add dependencies used by the RDT and ASF/RTP code

Originally committed as revision 22538 to svn://

12 years agoffmpeg: Combine variable declaration and definition
Alexander Strange [Mon, 15 Mar 2010 02:33:49 +0000 (02:33 +0000)]
ffmpeg: Combine variable declaration and definition

Originally committed as revision 22537 to svn://

12 years agoffmpeg: Factor out redundant sync_ipts calculation
Alexander Strange [Mon, 15 Mar 2010 02:32:21 +0000 (02:32 +0000)]
ffmpeg: Factor out redundant sync_ipts calculation

Originally committed as revision 22536 to svn://

12 years agomatroskadec: use more appropriate error code
Aurelien Jacobs [Mon, 15 Mar 2010 00:19:22 +0000 (00:19 +0000)]
matroskadec: use more appropriate error code

Originally committed as revision 22535 to svn://

12 years agomove ff_url_split() and ff_url_join() declarations to internal.h
Aurelien Jacobs [Sun, 14 Mar 2010 23:59:48 +0000 (23:59 +0000)]
move ff_url_split() and ff_url_join() declarations to internal.h
those functions are not part of the public API

Originally committed as revision 22534 to svn://

12 years agoCosmetics: Fix a comment.
Carl Eugen Hoyos [Sun, 14 Mar 2010 23:52:45 +0000 (23:52 +0000)]
Cosmetics: Fix a comment.

Originally committed as revision 22533 to svn://

12 years agoMove the probe loop from av_open_input_file() into its own method
Micah F. Galizia [Sun, 14 Mar 2010 22:40:16 +0000 (22:40 +0000)]
Move the probe loop from av_open_input_file() into its own method
av_probe_input_buffer() so that it can be reused. Here are a few
differences to the original way things were probed:

- maximum probe buffer size can be specified as a parameter.

- offset within the stream to probe from can be specified as a parameter.

- instead of seeking back to the start each time a probe fails, stream
  data is appended to the reallocated buffer. This lowers the amount
  of data read from the stream (there is no repetition) and results in
  fewer closed and reopened streams (when seeking fails).

New attempt after r22296, which was revert in r22315 due to a FATE

See the thread:
Subject: [FFmpeg-devel] [PATCH] Move av_open_input_file probe loop to its own method
Date: 2010-03-05 03:23:57 GMT

Patch by Micah F. Galizia printf("%s%s@%s.%s", "micah", "galizia", "gmail", "com").

Originally committed as revision 22532 to svn://

12 years agoDWT: x86 init should depend on HAVE_MMX
Måns Rullgård [Sun, 14 Mar 2010 22:29:11 +0000 (22:29 +0000)]
DWT: x86 init should depend on HAVE_MMX

The init function is only compiled if MMX is enabled, the call
must use the same condition.

Originally committed as revision 22531 to svn://

12 years agoerror.h: test EDOM instead of EINVAL
Måns Rullgård [Sun, 14 Mar 2010 22:25:35 +0000 (22:25 +0000)]
error.h: test EDOM instead of EINVAL

C99 doesn't require EINVAL, only EDOM, EILSEQ, and ERANGE.

Originally committed as revision 22530 to svn://

12 years agoAdd missing includes to libavutil/error.h
Måns Rullgård [Sun, 14 Mar 2010 22:25:31 +0000 (22:25 +0000)]
Add missing includes to libavutil/error.h

Originally committed as revision 22529 to svn://

12 years agoChange the definition of AVERROR_NUMEXPECTED at the next libavutil
Stefano Sabatini [Sun, 14 Mar 2010 21:16:29 +0000 (21:16 +0000)]
Change the definition of AVERROR_NUMEXPECTED at the next libavutil
major bump, using an FFmpeg specific error code rather than EDOM,
which has a quite different semantics.

Originally committed as revision 22528 to svn://

12 years agoAllow hard-coding of the 32kB cubic-root table for AAC.
Reimar Döffinger [Sun, 14 Mar 2010 19:59:47 +0000 (19:59 +0000)]
Allow hard-coding of the 32kB cubic-root table for AAC.

Originally committed as revision 22527 to svn://

12 years agoFix brief make output for generated tables
Måns Rullgård [Sun, 14 Mar 2010 19:48:09 +0000 (19:48 +0000)]
Fix brief make output for generated tables

Originally committed as revision 22526 to svn://

12 years agoAllow hard-coding several QDM2 tables (about 32 kB size).
Reimar Döffinger [Sun, 14 Mar 2010 19:30:25 +0000 (19:30 +0000)]
Allow hard-coding several QDM2 tables (about 32 kB size).

Originally committed as revision 22525 to svn://

12 years agoAdd some more table-printing functions needed for future patches.
Reimar Döffinger [Sun, 14 Mar 2010 19:27:58 +0000 (19:27 +0000)]
Add some more table-printing functions needed for future patches.

Originally committed as revision 22524 to svn://

12 years agoAdd ff_ prefix to dwt functions
Måns Rullgård [Sun, 14 Mar 2010 17:50:16 +0000 (17:50 +0000)]
Add ff_ prefix to dwt functions

Originally committed as revision 22523 to svn://

12 years agoSeparate DWT from snow and dsputil
Måns Rullgård [Sun, 14 Mar 2010 17:50:12 +0000 (17:50 +0000)]
Separate DWT from snow and dsputil

This moves the DWT functions from snow.c and dsputil.c to a file of
their own.  A new struct, DWTContext, holds the function pointers
previously part of DSPContext.

Originally committed as revision 22522 to svn://

12 years agoAvoid creating tiny (possibly only 64 bytes large) audio packets resulting in
Reimar Döffinger [Sun, 14 Mar 2010 13:38:00 +0000 (13:38 +0000)]
Avoid creating tiny (possibly only 64 bytes large) audio packets resulting in
huge processing and memory usage overhead for avi files with raw PCM audio.

Originally committed as revision 22521 to svn://

12 years agoMake sure all mvs of a mb are set in the error concealment code.
Michael Niedermayer [Sun, 14 Mar 2010 01:52:31 +0000 (01:52 +0000)]
Make sure all mvs of a mb are set in the error concealment code.

Originally committed as revision 22520 to svn://

12 years agoMark AVERROR_ENOENT for deletion at the next libavutil major bump.
Stefano Sabatini [Sun, 14 Mar 2010 01:32:44 +0000 (01:32 +0000)]
Mark AVERROR_ENOENT for deletion at the next libavutil major bump.

The symbol is currently unused, AVERROR(ENOENT) must be used instead.

Originally committed as revision 22519 to svn://

12 years agoLexically sort the error code definitions.
Stefano Sabatini [Sun, 14 Mar 2010 01:32:41 +0000 (01:32 +0000)]
Lexically sort the error code definitions.

Originally committed as revision 22518 to svn://

12 years agoEnsure that the deblock filter accesses the correct MVs for h264.
Michael Niedermayer [Sun, 14 Mar 2010 00:42:26 +0000 (00:42 +0000)]
Ensure that the deblock filter accesses the correct MVs for h264.

Originally committed as revision 22517 to svn://

12 years agoaacsbr: DSPUtilize analysis input scaling.
Alex Converse [Sat, 13 Mar 2010 21:12:45 +0000 (21:12 +0000)]
aacsbr: DSPUtilize analysis input scaling.

Originally committed as revision 22516 to svn://

12 years agoaac: Don't bias output during the IMDCT if SBR will be used.
Alex Converse [Sat, 13 Mar 2010 21:12:10 +0000 (21:12 +0000)]
aac: Don't bias output during the IMDCT if SBR will be used.

Originally committed as revision 22515 to svn://

12 years agoMark AVERROR_ENOMEM to be deleted at the next major bump.
Stefano Sabatini [Sat, 13 Mar 2010 19:19:46 +0000 (19:19 +0000)]
Mark AVERROR_ENOMEM to be deleted at the next major bump.

AVERROR(ENOMEM) must be used instead, and there are no occurrences of
AVERROR_ENOMEM in the FFmpeg basecode so it can be safely dropped.

Originally committed as revision 22514 to svn://

12 years agoReplace all the occurrences of AVERROR_EIO with AVERROR(EIO), and mark
Stefano Sabatini [Sat, 13 Mar 2010 19:19:44 +0000 (19:19 +0000)]
Replace all the occurrences of AVERROR_EIO with AVERROR(EIO), and mark
AVERROR_EIO for deletion at the next major bump.

Originally committed as revision 22513 to svn://

12 years agoMake iff.c:decode_init return the value returned by
Stefano Sabatini [Sat, 13 Mar 2010 19:19:40 +0000 (19:19 +0000)]
Make iff.c:decode_init return the value returned by
avctx->get_buffer() in case of error, rather than AVERROR_UNKNOWN
which is deprecated, and mark AVERROR_UNKNOWN for deletion at the next
major bump.

Originally committed as revision 22512 to svn://

12 years agoCODEC_ID_XVID is not used anywhere, remove it at next major bump
Aurelien Jacobs [Sat, 13 Mar 2010 18:19:36 +0000 (18:19 +0000)]
CODEC_ID_XVID is not used anywhere, remove it at next major bump

Originally committed as revision 22511 to svn://

12 years agobump minor avcodec version for libxvid codec id change (r22489)
Aurelien Jacobs [Sat, 13 Mar 2010 18:17:25 +0000 (18:17 +0000)]
bump minor avcodec version for libxvid codec id change (r22489)

Originally committed as revision 22510 to svn://

12 years agoMake the scale filter pass the pkt.pos information to the output
Stefano Sabatini [Sat, 13 Mar 2010 12:33:44 +0000 (12:33 +0000)]
Make the scale filter pass the pkt.pos information to the output

Originally committed as revision 22509 to svn://

12 years agoBump libavfilter minor version and add APIchanges entry after the
Stefano Sabatini [Sat, 13 Mar 2010 11:41:15 +0000 (11:41 +0000)]
Bump libavfilter minor version and add APIchanges entry after the
addition of the AVFilterPicRef.pos field.

Originally committed as revision 22508 to svn://

12 years agoMake ffplay pass the packet.pos information to the filterchain, get it
Stefano Sabatini [Sat, 13 Mar 2010 11:27:07 +0000 (11:27 +0000)]
Make ffplay pass the packet.pos information to the filterchain, get it
again from the filterchain and pass it back to output_picture2().

Originally committed as revision 22507 to svn://

12 years agoAdd AVFilterPicRef.pos field and make libavfilter propagate stream
Stefano Sabatini [Sat, 13 Mar 2010 11:27:05 +0000 (11:27 +0000)]
Add AVFilterPicRef.pos field and make libavfilter propagate stream
byte position information, as stored in the pkt.pos, through the

Note that the pos field is added *non* at the end of the
AVFilterPicRef struct, thus breaking ABI compatibility, which is
allowed as the API is still considered non-stable.

Originally committed as revision 22506 to svn://

12 years agoExtend the slice filter to make it issue slice height values randomly
Stefano Sabatini [Sat, 13 Mar 2010 10:41:25 +0000 (10:41 +0000)]
Extend the slice filter to make it issue slice height values randomly
choosen between 8 and 32 when the supplied parameter is the string

This is useful for testing the slice support, but it is not supposed
to be used for other purposes and this interface may change in the
future, thus it is not documented.

The randomization algorithm adopted is the standard Numerical Recipes

Originally committed as revision 22505 to svn://

12 years agovp3: Split out motion vectors to their own array
David Conrad [Sat, 13 Mar 2010 10:25:41 +0000 (10:25 +0000)]
vp3: Split out motion vectors to their own array

1.5% faster overall decode on my penryn

Originally committed as revision 22504 to svn://

12 years agoAdd APIchanges entry for the libavutil/error.h addition.
Stefano Sabatini [Sat, 13 Mar 2010 09:59:36 +0000 (09:59 +0000)]
Add APIchanges entry for the libavutil/error.h addition.

Originally committed as revision 22503 to svn://

12 years agoAdd missing parentheses around the AVERROR_PATCHWELCOME macro
Stefano Sabatini [Sat, 13 Mar 2010 09:52:43 +0000 (09:52 +0000)]
Add missing parentheses around the AVERROR_PATCHWELCOME macro

Originally committed as revision 22502 to svn://