Re: Theora packets with granulepos of -1

>> The value of the first keyframe is irrelevant for seeking. The corner
>
> This is not true. You have to know whether the logical frame index (the
> frame count after you add together the two pieces of the split granpos)
> starts counting from 0 or 1 in order to map it to a time correctly.

Oh, I read the original as 'whatever offset the stream starts at'...