Hello

I am implementing speech recognizer using hidden markov model.First 
step in implementing HMM is to find model parameters. 

I dont know to devide my speech signal into states. I am planing to do 
implement HMM with 5 states. so do i need to devide length of the 
sample by the number of states and assigned to each state??

Should number of the observed symbols in each state be the same??

I read somewhere that each trail was split into frames. does it mean 
each frame is one observed symbol??

any kind of help or suggestions are welcome.thanks for your time.

Thanks
Nivas

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