[theora-dev] Theora packets with granulepos of -1

ogg.k.ogg.k at googlemail.com ogg.k.ogg.k at googlemail.com
Mon Feb 16 05:18:17 PST 2009

> Well I understand that codecs used in an ogg container are lossy like
> Theora, Speex and Vorbis are. But why making the important basement - the
> container itself - lossy? Doesn't make any sense to me especially related to
> the granule position because this part is not compressed or smaller because
> a -1 is used instead of the right value.

The -1 is implicit: there is one granpos per page, and all the packets ending
on that page get their "virtual' granpos deduced from this page granpos
and the length of the packet (eg, 1/fps for theora, samples*rate for vorbis).

> And not to forget: This important issue is nowhere mentioned (as far as I
> kow) especially when it comes to seeking. Why is there no addendum to the
> Theora specs to discuss the issue of -1 packets if used in the ogg
> container? I think this is quite important. There is even a chapter about
> Theora in ogg container.

Most seeking is done on pages.

> Are there also similar issues with vorbis? Are there also -1 packets to make
> the packet-handling... uhm, more funny?

Yes, see above.

More information about the theora-dev mailing list