Bug introduced by:
commit f12f40b31a
Author: Anton Khirnov <anton@khirnov.net>
Date: Tue Aug 30 04:05:20 2011 +0200
ffmpeg: get rid of new* options.
They are confusing, irregular and redundant -- -map already contains all
the information. Stream maps can now be parsed in opt_output_file().
Add a more user-friendly default behavior in case no maps are present.
Breaks -programid for now, but it never worked properly anyway. A better
solution will be written soon.
Bug has been introduced by:
commit 1cede1d011
Author: Anton Khirnov <anton@khirnov.net>
Date: Sat Jun 25 07:43:49 2011 +0200
ffmpeg: factor common code from new_a/v/s/d_stream to new_output_stream()
Bug introduced by:
commit e3bf4e292c
Author: Anton Khirnov <anton@khirnov.net>
Date: Tue Aug 30 04:10:54 2011 +0200
ffmpeg: replace -vcodec/-acodec/-scodec with a better system.
The new option doesn't depend on its placement wrt -new* options (which
don't exist anymore) and works in a similar way as per-stream AVOptions.
-[vas]codec remain as aliases to -codec:[vas]
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Now that the option was moved in the per-stream context,
the parsing is done before the time_base for the stream is decided.
This patch does the parsing in AV_TIME_BASE units and rescales the
timestamps later when the correct time base is known.
Signed-off-by: Nicolas George <nicolas.george@normalesup.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Also mark with a visible comment "FIXME realloc failure" places where
av_realloc seems to lack a proper test for failure.
Signed-off-by: Nicolas George <nicolas.george@normalesup.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This reverts commit 2cf8355f98.
AVInputStream.nb_streams tracks number of streams found at the
beginning, new streams may appear that ffmpeg doesn't know about. Fixes
crash in this case.
Similar to libswscale this does resampling and format convertion, just for audio
instead of video.
changing sampling rate, sample formats, channel layouts and sample packing all
in one with a very simple public interface.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This is done in order to clarify the non-video-specific nature of the
buffersink code, as the result of the video/audio API unification of
the previous commit, and for improving overall consistency.
The new API is more generic (no distinction between audio/video for
pulling frames), and avoids code duplication.
A backward compatibility layer is kept for avoiding tools ABI breaks
(only for the video binary interface, audio interface was never used
in the tools).