The mov/mp4 muxer has support for handling negative timestamps
via edit lists (which customarily is used for handling the 1-frame
delay due to B-frames as well).
Using the muxer's native way of handling it is better than using
the generic offsetting. The generic offsetting is a bit too
crude when e.g. the timebase of one track is 1/fps, where the
edit lists can handle it accurately.
Signed-off-by: Martin Storsjö <martin@martin.st>
This reverts commit bf36dc50ea448999c8f8c7a35f6139a7040f6275, reversing
changes made to b7fc2693c70fe72936e4ce124c802ac23857c476.
Conflicts:
libavcodec/h264.c
Keeping support for the old VDPAU API has been requested by our VDPAU maintainer
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Uses the 2.5 compatibility header included with the variant of
FFMS2 that uses AviSynth's C-interface. A copy of this header is
now provided in compat/avisynth.
avs_get_row_size_p and avs_get_height_p changed between versions
2.5 and 2.6. Since the avisynth_c.h header that avformat uses
assumes AviSynth 2.6, it would cause 2.5 to crash if given any
kind of real video (the Version() function was known to work,
though).
AvxSynth was unaffected by this issue because, despite being based
on AviSynth 2.5.8 and using 2.5.8's interface version number of 3,
it actually uses 2.6's versions of these functions.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The case where decoding should only happen for keyframes and afterwards
(CODEC_FLAG2_SHOW_ALL not set) does not fully work neither before nor
after this
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master:
dxa: Make sure the reference frame exists
See: 12eb2fd5394813a6119912b22f5dfc17b4a6b4a0
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'c59967fa7cc5bc2fa06b36c17d2c207240c06b3e':
h261: check the mtype index
Conflicts:
libavcodec/h261dec.c
The right half of the check is impossible to be true and thus replaced by
an assert()
See: ec3cd74f2dab8e3e8234ccb994132b23d3098585
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '9d86bfc259ae9ba7a76067ec931ff20fbb86ea2a':
hlsenc: Don't reset the number variable when wrapping
Conflicts:
libavformat/hlsenc.c
See: 75b9fb27f516f9db7995ab2c2abb83e25cae5813
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '09c93b1b957f2049ea5fd8fb0e6f4d82680172f2':
hlsenc: Append the last incomplete segment when closing the output
Conflicts:
libavformat/hlsenc.c
See: 4630dfd1ebb6550f352ea69ce7150aaaba96d413
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'c9031c7c1446a1a63eff7c0bf50d1ee559adf3fb':
hlsenc: Add a proper dependency on the mpegts muxer
Conflicts:
libavformat/Makefile
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'cb214707a6cb0d3272ec0261af6f1f5d8b7dabc7':
vp56data: Move all shared enum/struct declarations to common header
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '38f64c03301ac66d7b54b3e4bd2bf6454f9fb2d3':
mpeg12decdata.h: Move all tables to the only place they are used
Conflicts:
libavcodec/mpeg12decdata.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'c591d4575a6f97fbbe6145304b1ea960e8e81e14':
avcodec: Replace local extern declarations for tables with header #includes
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'd258531502b24cb653204fe4f003c8815755bdc4':
swscale: Mark a bunch of tables only used within one file static
Conflicts:
libswscale/swscale_unscaled.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '060ce0c697e261ca2792a7df30dfd1bae6900a4f':
ivi_common: Make some tables only used within the file static
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '4b054a3400f728c54470ee6a1eefe1d82420f6a2':
rtpproto: Check the right feature detection macro
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '163a729725c6eb0081b0af41a7279f7d19aee86e':
electronicarts: Let functions always returning the same value return void
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Implicit vector loads on POWER7 hardware can use the VSX
instruction set instead of classic Altivec/VMX. Let's force
a VMX load in this case.
Signed-off-by: Martin Storsjö <martin@martin.st>
The counter itself shouldn't be wrapped, since it is used for
determining end_pts for the next segment - only wrap the number
used for the segment file name.
Signed-off-by: Martin Storsjö <martin@martin.st>
The hls muxer itself doesn't have any direct (object file level)
dependencies on mpegtsenc.o, and including that object file
directly doesn't ensure that it is registered so that the muxer
actually is accessible.
Signed-off-by: Martin Storsjö <martin@martin.st>
This allows specifying more dither algorithms without using up flags and
without ambiguities.
Also initialize the new field based on the flags and use it.
Note, improving the logic of the checks is left to subsequent
commits, this here only switches from flags to enum.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
