Hi

  Im working on implementation of a very basic speech recognizer 
using TIMIT Database( right now with just one speaker) in MATLAB.
1) How do I process the audio files(each speaker has 10 files) using 
waveread. Do I have to use a for loop?

2) What do I take the value of K as in the k-means clustering?

3) After K-means comes the HMM training right? So will the input 
sequence to hmmtrain(matlab function) be the output of K-means 
clustering as it is? What should the initial estimate of the 
transition and emission probabilities be? 

4) How do I build a HMM for each phoneme of the speaker's speech? Ho 
do I obtain the phoneme in the first place?

I'd appreciate your help. Thank you

Namrata

(You need to be a member of speech-recognition -- send a blank email to speech-recognition-subscribe@yahoogroups.com )