ogg dev - Jun 2009 - Ogg Skeleton

Hi,

I got a short question regarding the ogg skeleton.

As far as I understand it, the skeleton information helps to decode the timeing
from the granule position, that is available in every ogg page, so that I can
say, what is the end time of the last full ogg packet within an ogg page.

So but it does not help me finding out the timing information about the other
packets in between?

-Yorn

> As far as I understand it, the skeleton information helps to decode the
> timeing from the granule position, that is available in every ogg page, so
It holds the granulerate, so it'd allow this for those codecs that use the
granule shift way of encoding their granules (except for those pages
where no packet ends, as these will have no granpos set).
> that I can say, what is the end time of the last full ogg packet within an
> ogg page.
Actually, not necessarily.For some codecs, it's the start time, and that
is OK by Ogg rules. The granpos of a page is (I _think_) that of the last
packet (if any) ending on that page. It is *mostly* true though.
> So but it does not help me finding out the timing information about the
> other packets in between?
No. You have to know about the codec, which can reconstruct those.
Typically, if no automatic reconstruction is possible, you'll have the
codec flush after each packet (that's what I do).

On Fri, Jun 5, 2009 at 4:30 AM, ogg.k.ogg.k at googlemail.com
<ogg.k.ogg.k at googlemail.com> wrote:
> It holds the granulerate, so it'd allow this for those codecs that use
the
> granule shift way of encoding their granules (except for those pages
> where no packet ends, as these will have no granpos set).
BTW, we need to add another field out of the remaining reserved bits
to accommodate reordering codecs like dirac.

 -r