Re: seeking in time, feedback+question

> I do not know how you could get rid of the problem,
> but you could mask it by 
> starting the decode a few frames early, and throwing
> away the audio for 
> those early frames.  Then the decode state will have
> converged when you 
> start decoding the audio that you want.
> 
>  Regards,
> 
> Jim Crichton 

Thanks!
I exactly do that but the problem is that I need to
know "how much" to read ahead and throw away since I
want to use it in an application on the fly. Actually
my observation is that for unvoiced frames it catches
up quite quickly but for voiced frames I may need to
decode up to 20 frames(frameSize=320) which is a lot.
I guess it is due to the residual history and for
unvoiced frames consequitive frames depend on each
other less compared to voiced frames. For the moment I
am thinking of starting decoding at the first previous
unvoiced frame(I'll use my own voiced-unvoiced
decision for that).
Regards,
Baris

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com