Acoustic Model Discussions

Flat
Re: Sphinx4 implementation
User: Moiz
Date: 8/27/2009 12:09 am
Views: 126
Rating: 1

Dear Sir i have uploaded my files on thins link. please check it out

https://ssl.rapidshare.com/cgi-bin/collectorszone.cgi?

Guide me where i m wrong.

 


Waiting for your Replyyyyyyyyyyyyy


Thanks & Regards

 

 

--- (Edited on 8/27/2009 12:09 am [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: kmaclean
Date: 8/27/2009 11:10 am
Views: 45
Rating: 1

Hi Moiz,

>https://ssl.rapidshare.com/cgi-bin/collectorszone.cgi?

I don't hink you posted the full link to your file...

Ken

--- (Edited on 8/27/2009 12:10 pm [GMT-0400] by kmaclean) ---

Re: Sphinx4 implementation
User: moiz
Date: 8/27/2009 10:06 pm
Views: 47
Rating: 1

Here is the full link... sorry for the inconvinency...

https://ssl.rapidshare.com/cgi-bin/collectorszone.cgiSmile

--- (Edited on 8/27/2009 10:06 pm [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: Moiz
Date: 8/27/2009 10:22 pm
Views: 53
Rating: 1

Here is another link. please check it out.

http://www.4shared.com/file/128278186/a870edb7/sphinxtraintar.html

Thank

 

--- (Edited on 8/27/2009 10:22 pm [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: nsh
Date: 8/28/2009 1:30 am
Views: 37
Rating: 1

Your database is ridiculously small. You need 200 times (an hour of recordings) more data in order to recognize something. Also in the configuration file you need to use smaller amount of tied states:

$CFG_N_TIED_STATES = 250;

--- (Edited on 8/28/2009 1:30 am [GMT-0500] by nsh) ---

Re: Sphinx4 implementation
User: Moiz
Date: 8/28/2009 3:07 am
Views: 130
Rating: 1

Thankyou Very Much.

I will try it out.

--- (Edited on 8/28/2009 3:07 am [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: Moiz
Date: 8/29/2009 11:57 am
Views: 112
Rating: 1

Hello Sir,

            can u provide me the perfect step to train accoustic model perfectly and how to use that train accoutic model in sphinx4.

any link or what ever.

Thanks in advance.

--- (Edited on 8/29/2009 11:57 am [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: coarline
Date: 8/31/2009 3:31 am
Views: 124
Rating: 1

Hi,  Is speech recognition possible with small data. if yes how should i train that data and implement it in sphinx4 to be recognized perfectly. please suggest me some resources.

thanks.

Coarline.

--- (Edited on 8/31/2009 3:32 am [GMT-0500] by Visitor) ---

Re: Sphinx4 implementation
User: nsh
Date: 9/1/2009 4:14 am
Views: 2506
Rating: 1

Dear Coarline and others


Everything depends on details. If you really need help provide more information - who are you, what are you trying to do, what have you done already and what problems do you have right now.

It's perfectly possible to build a high-accuracy recognition system with sphinx4. Adaptation of the exitsting models helps with the sparsity of the data. But there is no general reciept on how to reach required accuracy.

 

--- (Edited on 9/1/2009 4:14 am [GMT-0500] by nsh) ---

PreviousNext