should make Nathan Kurz and if I don't hear too much complaints
about it -- that's the way it will be from now on.
* updating regressions
Originally committed as revision 3527 to svn://svn.ffmpeg.org/ffmpeg/trunk
finetune bit/distortion weighting factor used in motion estimation, the old coeffs where finetuned relative to incorrect mv_penalty tables which where then fixed later but the coeffs where not
this _may_ fix the long standing blocking artifacts, but may also introduce mudding artefacts theoretically, so please tell us if u stumble across any so we can either fix them or export this variable so the user can change it
Originally committed as revision 3189 to svn://svn.ffmpeg.org/ffmpeg/trunk
dts/pts fixed for streamcopy
dont use coded_frame->key_frame hack in muxers, use AVPacket.flags instead
Originally committed as revision 3171 to svn://svn.ffmpeg.org/ffmpeg/trunk
fixes the random dts/pts during encoding
asf preroll fix
no more initial zero frames for b frame encoding
mpeg-es dts during demuxing fixed
.ffm timestamp scale fixed, ffm is still broken though
Originally committed as revision 3168 to svn://svn.ffmpeg.org/ffmpeg/trunk
- fixed VBR+constrained bitstream header flags for non-VCD
- more sane (and SVCD compatible) value for video stream->max_buffer_size
- always write at least one PES header stuffing byte for MPEG-2 to
prevent accidental start code generation
- do not write more than 16 stuffing bytes in a PES header (not
allowed). Use padding packets instead.
- include a PES extension in the first MPEG-2 packet
- fill the first pack of SVCD files with padding
- "sanity hack" that prevents the SCR from overtaking the PTS for non-VCD
- fixed VCD PTS values to correspond to the SCR
- always include DTS in the first SVCD packet (fixes lots of
compatibility problems with DVD players)
patch by (Hauke Duden <H.NS.Duden at gmx dot net>)
Originally committed as revision 3078 to svn://svn.ffmpeg.org/ffmpeg/trunk
1. Corrected packet sent time. (Packet sent time => previous packet sent
time)
2. Corrected payload playing time. (Payload playing time => packet sent
time)
3. Used different format for packets with one payload and multi payloads.
Originally committed as revision 2851 to svn://svn.ffmpeg.org/ffmpeg/trunk
* adding regression tests for all pix_fmt conversions
* making libavtest part of the default testsuite
Originally committed as revision 1855 to svn://svn.ffmpeg.org/ffmpeg/trunk
Michael Niedermayer
Roman Shaposhnik
Hauke Duden
Steven M. Schultz
Konstantin Andreyev
Fabrice Bellard