Acoustic Model Discussions

Flat
Creating user independant models
User: bedahr
Date: 2/21/2010 3:01 pm
Views: 4910
Rating: 2

Hello!

 

Ok this is probably a stupid question but I'll ask anyways. 

 

I was wondering what the difference is when creating speaker independant speech models instead of speaker dependant models - apart from using samples from multiple persons of course.

 

I looked around in the dev/ page but I couldn't find it: Is the script used to create the Voxforge acoustic models available somewhere?

 

Greetings,

Peter

--- (Edited on 2/21/2010 3:01 pm [GMT-0600] by bedahr) ---

Re: Creating user independant models
User: kmaclean
Date: 2/21/2010 8:00 pm
Views: 57
Rating: 2

>I was wondering what the difference is when creating speaker independant

>speech models instead of speaker dependant models - apart from using

>samples from multiple persons of course.

There is no difference for the HTK acoustic models I create for VoxForge.  I don't know if there should be a difference.

>I looked around in the dev/ page but I couldn't find it: Is the script used to

>create the Voxforge acoustic models available somewhere?

AcousticModel.pm calls Compile_AcousticModel.sh

Basically the same approach used in the VoxForge Howto/Tutorial.

Ken

 

--- (Edited on 2/21/2010 9:00 pm [GMT-0500] by kmaclean) ---

Re: Creating user independant models
User: bedahr
Date: 2/22/2010 3:42 pm
Views: 2151
Rating: 2

Thanks for the quick and very helpful response.

 

Greetings,

Peter

--- (Edited on 2/22/2010 3:42 pm [GMT-0600] by bedahr) ---

PreviousNext