Hello,
In the mail archives from 2004 I found this:
> Note that this (MCRA and all) is only true for the "new" VAD
included in the
> 1.1.x preprocessor. The other (main) VAD is just a big kludge that could
> certainly be improved.
>
> Jean-Marc
>
> Selon Devendra Jalihal <dj at ee.iitm.ac.in
<http://lists.xiph.org/mailman/listinfo/speex-dev>>:
>
> > Many of the ideas used in Speex VAD are found in the paper by Israel
> > Cohen,
> > "Noise Spectrum Estimation in Adverse Environments: Improved
Minima
> > Controlled Recursive Averaging"
> > that appeared in IEEE Trans on Speech and Audio Processing, Sept 2003.
> > -Devendra Jalihal
*Is VAD in Speex 1.2rc1 based on MCRA or does it use something else?*
Thanks,
Goran
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20090918/6e71eacf/attachment.htm