Prev: normalization
Next: Comprression ratio for SPIHT algorithm
From: Joshua on 4 Mar 2010 01:01
serdar <serdarkurcan(a)yahoo.com> wrote in message <ef17838.-1(a)webx.raydaftYaTP>...
> Hi everybody,
>
> I am working on my MSEE thesis about a speech recognizer. I was
> wondering anybody out there has a Matlab code or a working algorithm
> for endpoint detection of isolated utterances in speech.
>
> I am currently trying the zero-crossing and the short-term energy
> plots to distinguish the endpoints of an isolated word from noise
> and/or silence part of speech.
>
> Does anybody know a better or more effective way of working out the
> endpoint detection problem ? I would appreciate any working code or
> algorithm on this problem.
>
> Thanks.

I would look into wavelet decomposition combine with energy...

just a thought...
 | 
Pages: 1
Prev: normalization
Next: Comprression ratio for SPIHT algorithm