[Speex-dev] Speech Detection

[Zhao, Dun (NSN - CN/Shanghai)]

Hi Debasish,
 
I think it is a better way to refer speech recognition materials.
The preprocess part of them will always introduce the words detection.
 
One way I used is to compute the zero-crossing rate.
 
Here is the wiki page you may find useful:
http://en.wikipedia.org/wiki/Zero-crossing_rate
 
 
Good luck!
 
 
Martin

________________________________

From: speex-dev-bounces at xiph.org [mailto:speex-dev-bounces at xiph.org] On
Behalf Of ext Debasish Chandra
Sent: Wednesday, January 06, 2010 12:23 AM
To: speex-dev at xiph.org
Subject: [Speex-dev] Speech Detection



Hi,

 

I want to detect speech locations from raw WAV file. For example a 10
sec WAV file contains silence/noise for 1st sec, speech for next 4
seconds, then silence/noise for next 2 seconds and speech for next 3
seconds. 

I want to generate an information that this WAV file has speech from 1-5
seconds and from 8-10 seconds.  

How I can do that by using preprocessor utility of libspeexdsp?

 

With Best Regards,

Debasish

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/speex-dev/attachments/20100106/e5c1ef7b/attachment.htm