Jason Self
2009-Nov-04 04:07 UTC
[theora] Better A/V sync and framerate detection in ffmpeg2theora
In addition to the indexing mentioned earlier by Chris Pearce, improvements to frame rate detection and A/V sync were recently committed to ffmpeg2theora. This will help with variable framerate video and in other cases where the container specifies the incorrect framerate. Remember those cases where it would report the framerate as double what it really was? That seems to be fixed now. In my testing, ffmpeg2theora should now keep lip-synch as well as any DVD player can. https://trac.xiph.org/changeset/16650
Andrew Chew
2009-Nov-04 04:11 UTC
[theora] Better A/V sync and framerate detection in ffmpeg2theora
On Nov 3, 2009, at 8:07 PM, Jason Self wrote:> In addition to the indexing mentioned earlier by Chris Pearce, > improvements to frame rate detection and A/V sync were recently > committed to ffmpeg2theora. This will help with variable framerate > video and in other cases where the container specifies the incorrect > framerate. > > Remember those cases where it would report the framerate as double > what it really was? That seems to be fixed now. In my testing, > ffmpeg2theora should now keep lip-synch as well as any DVD player can. > > https://trac.xiph.org/changeset/16650Sweet! Is a new official ffmpeg2theora build (.26) imminent?
Christopher Blizzard
2009-Nov-04 07:58 UTC
[theora] Better A/V sync and framerate detection in ffmpeg2theora
Also wondering if there are any fixes for the encoder that might make for a good 1.1.1 release. (Getting to a cadence of releases every few weeks makes a big difference over the long term.) --Chris
j at v2v.cc
2009-Nov-04 10:55 UTC
[theora] Better A/V sync and framerate detection in ffmpeg2theora
On 11/04/2009 05:11 AM, Andrew Chew wrote:> > On Nov 3, 2009, at 8:07 PM, Jason Self wrote: > >> In addition to the indexing mentioned earlier by Chris Pearce, >> improvements to frame rate detection and A/V sync were recently >> committed to ffmpeg2theora. This will help with variable framerate >> video and in other cases where the container specifies the incorrect >> framerate. >> >> Remember those cases where it would report the framerate as double >> what it really was? That seems to be fixed now. In my testing, >> ffmpeg2theora should now keep lip-synch as well as any DVD player can. >> >> https://trac.xiph.org/changeset/16650 > > Sweet! Is a new official ffmpeg2theora build (.26) imminent?once indexing and seeking got some more testing i will prepare a new release. j
Jason Self
2009-Nov-04 17:31 UTC
[theora] Better A/V sync and framerate detection in ffmpeg2theora
On Nov 3, 2009, at 11:58 PM, Christopher Blizzard wrote:> Also wondering if there are any fixes for the encoder that might > make for a good 1.1.1 release. (Getting to a cadence of releases > every few weeks makes a big difference over the long term.)libtheora already has a 1.1.1 release: http://lists.xiph.org/pipermail/theora/2009-October/002827.html Time-based releases might be an interesting idea, but you can already compile from svn whenever you want to get the latest changes.