[Speex-dev] VAD levels and offsets.

[taintedham-mailinglists at yahoo.com]

Hello everyone,
I would like to segment a file by speech and silence. 
I know that I can change some params in speex to
change the prob. of detection.

What I would like to do... is be able to while
decoding, or during encoding....

Read the state of the encoder/decoder to get back the
offsets for speech and silence.

Is there an existing utility for this?  If there isn't
could I quickly code something up?

Or is it even possible?

I figure, you guys have spent the time on the VAD
work, maybe I can benefit from that work!  Thanks!

--Dave