[theora] Ogg index and Skeleton 4.0

Philip Jägenstedt philip at foolip.org
Thu Apr 29 08:05:02 PDT 2010

On Thu, Apr 29, 2010 at 12:47, Chris Pearce <chris at pearce.org.nz> wrote:
> Hi everybody,
> I've updated OggIndex to now output Skeleton 4.0 tracks. The differences
> between Skeleton 3.x and Skeleton 4.0 with OggIndex is:
> The fisbone packet now includes a "Radix" field.
> The fisbone packet now includes two new compulsory message headers; "Role"
> and "Name".
> The fishead packet no longer includes "start time" and "end time" fields.
> The index packets for each track now include the "start time" and "end time"
> fields for that track. The duration can now be calculated as the greatest
> end time of all active tracks minus the smallest start time of all active
> tracks. Thanks to Benjamin for pointing out the need for this.
> I've removed the requirement in my spec that a keypoint must index the first
> keyframe in their page only, and reworded my validity conditions.
> The new spec can be found here:
> http://github.com/cpearce/OggIndex/blob/master/Skeleton-4.0-Index-Specification.txt
> If anybody thinks anything else should be added to Skeleton 4.0, please
> speak up now!
> I'm going to create a new wiki page for Skeleton 4.0 to replace the existing
> Skeleton 3.0 xiph wiki page, and I'll produce a patch for ffmpeg2theora to
> output Skeleton 4.0 in the next few days.
> I'll hopefully be landing support for this in Firefox trunk builds in the
> coming weeks.
> All the best,
> Chris P.

Would you happen to have a sample file already? Reportedly GStreamer's
oggdemux chokes on Skeleton tracks it doesn't understand, so I'd like
to see if that's easily fixable in Opera before files with Skeleton
index hits the web.

Philip Jägenstedt

More information about the theora mailing list