[Theora-dev] Re: Questions, MMX and co.

[Ralph Giles]

On Thu, Aug 26, 2004 at 08:42:52AM +0100, Adam D. Moss wrote:

> >Well, encoder output should be bit-for-bit identical, at least for the 
> >reference implementation. On the decoder one can make speed/quality 
> >tradeoffs.
> 
> Is this the right way around?  I thought the DEcoder had to be
> bit-for-bit identical because that's what VP3 assumes (precise
> decoding rules allow unbound GOP size, because of supposedly no
> cumulative error).

Yes, as you and Timothy pointed out, I was thinking about this backward. 
What I meant to say was that the decoder must behave reliably, so the 
encoder can take its in-loop behavior into account when both sides 
perform speed vs. quality trade-offs.

 -r