Acoustic Model Discussions

Flat
How different between feature with qulifier and no one?
User: Noah
Date: 1/3/2011 8:40 pm
Views: 4676
Rating: 1

Dear all.

I'm using HTK to extract speech feature (by HCopy).

I've some problem about it.

I want to known how different between speech feature with qulifier s like _D,_E (ex. MFCC_D_E) and no one (ex. MFCC).

 

Is more speech qulifier, will get more accuracy? Why?

Best.

--- (Edited on 1/3/2011 8:40 pm [GMT-0600] by Noah) ---

Re: How different between feature with qulifier and no one?
User: kmaclean
Date: 3/21/2011 12:57 am
Views: 2091
Rating: 2

>how different between speech feature with qulifier s like

>_D,_E (ex. MFCC_D_E) and no one (ex. MFCC).

see this post: MFCC_D_N_Z_0 format

>Is more speech qulifier, will get more accuracy? Why?

depends on many things... we get reasonable results using MFCC_D_N_Z_0

 

 

 

--- (Edited on 3/21/2011 1:57 am [GMT-0400] by kmaclean) ---

PreviousNext