theora dev - Feb 2009 - Theora packets with granulepos of -1

Hello,

I'm just totally confused. In my theora streams encoded using ffmpeg2theora
(but also when using my own encoder) I have packets with a granulepos of -1 so I
can't identify the packet during a seeking operation correctly. I can also
see those strange value when I just print the packet granulepos before sending
it to the Theora decoder.
I know why there are PAGES with granularpos of -1 (because no packet ends in the
page) but why would PACKETS have a granularpos of -1 ?
It seems arbitrary so it's independent from keyframes.

In the Theora specs you can read:
"In this way the granulepos field increased monotonically as required by
the Ogg format, but contains information necessary to efficiently find the
previous keyframe to continue decoding after a seek."

This means having anything other than a monotonically increasing granulepos
would break the specification.
So related to the the values I get the spec is broken, isn't it?

Or what's wrong here?

Sincerely,
Robin Siegemund, dp


-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/theora-dev/attachments/20090216/6dd4ab9f/attachment.htm

> I'm just totally confused. In my theora streams encoded using
ffmpeg2theora
> (but also when using my own encoder) I have packets with a granulepos of -1
Are you sure that there really are packets with -1 granpos ?
ogginfo or oggz info should report those if there really are.
Could it be possible that the code that pulls packets after
feeding pages doesn't properly check return codes, and
*thinks* there is a packet available when more data would
be needed for a full packet ?

On Tue, Feb 24, 2009 at 10:20 AM, Timothy B. Terriberry
<tterribe at email.unc.edu> wrote:
> salsaman wrote:
>> I check the first positive granulepos, not the first page. So it will
>> always come out with the correct value.
>
> Really? What happens when the first positive granpos is for frame 2897
> and says the previous keyframe was at frame 2980? Does that mean the
> stream starts from 0 or 1?

It means the stream starts at 2980.

Gabriel.