The developers of Sphinx have set up a excellent documentation:

http://cmusphinx.sourceforge.net/sphinx4/  (Creating and using Acoustic Model Package)

--- (Edited on 8/21/2008 11:49 am [GMT-0500] by Visitor) ---

Hi dasinen,

>I want the procedure for training the wav files so that it will increse the

>recognition accuracy

Do you mean you are looking for a procedure to adapt the VoxForge acoustic model to your voice?

Which speech recognition engine are you using?

We have instructions for creating an acoustic model for Julius on the VoxForge site, and CMU's Robust group has instructions for creating acoustic models for the Sphinx group speech recognition engines.

Ken 

--- (Edited on 8/23/2008 9:51 am [GMT-0400] by kmaclean) ---

I guess I should have read the subject line before replying... :)

You are looking for the Robust Group tutorial.

Ken

--- (Edited on 8/23/2008 9:52 am [GMT-0400] by kmaclean) ---

hi,

thanks for u r intrest for all who had send the replies

now i am trying to train the files using the link provided in the first reply,and i will put the result response soon!!

i previously tried to train my wav files (as they are giving 60% accuracy with out training) to improve the accuracy !!

i confused at one place where i can put my wav files during the training process or any other file need to be used for that in place of using wav files ?? help me on this

thanks

--- (Edited on 8/25/2008 5:19 am [GMT-0500] by Visitor) ---

Hi dasinen,

This post might help: Creating Sphinx Acoustic Models

Ken

--- (Edited on 8/25/2008 8:42 am [GMT-0400] by kmaclean) ---

hi kmaclean,

thanks for the given link i hope it will help me!!

one thing is i trained the rm1 database using the Robust Group Tutorial and then i created a model by following the  this link

D:\sphinx4src\sphinx4-1.0beta\doc\UsingSphinxTrainModels.html

finally i created the model and now my problem is while using this created model in HelloWorld example it showing the following error help me on this

Loading...
Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 12
        at edu.cmu.sphinx.linguist.acoustic.tiedstate.MixtureComponent.transform
Stats(MixtureComponent.java:244)
        at edu.cmu.sphinx.linguist.acoustic.tiedstate.MixtureComponent.precomput
eDistance(MixtureComponent.java:197)
        at edu.cmu.sphinx.linguist.acoustic.tiedstate.MixtureComponent.<init>(Mi
xtureComponent.java:93)
        at edu.cmu.sphinx.model.acoustic.TOY_8gau_13dCep_16k_40mel_130Hz_6800Hz.
ModelLoader.createSenonePool(ModelLoader.java:614)
        at edu.cmu.sphinx.model.acoustic.TOY_8gau_13dCep_16k_40mel_130Hz_6800Hz.
ModelLoader.loadModelFiles(ModelLoader.java:558)
        at edu.cmu.sphinx.model.acoustic.TOY_8gau_13dCep_16k_40mel_130Hz_6800Hz.
ModelLoader.load(ModelLoader.java:476)
        at edu.cmu.sphinx.model.acoustic.TOY_8gau_13dCep_16k_40mel_130Hz_6800Hz.
Model.allocate(Model.java:177)
        at edu.cmu.sphinx.linguist.flat.FlatLinguist.allocateAcousticModel(FlatL
inguist.java:336)
        at edu.cmu.sphinx.linguist.flat.FlatLinguist.allocate(FlatLinguist.java:
318)
        at edu.cmu.sphinx.decoder.search.SimpleBreadthFirstSearchManager.allocat
e(SimpleBreadthFirstSearchManager.java:602)
        at edu.cmu.sphinx.decoder.Decoder.allocate(Decoder.java:109)
        at edu.cmu.sphinx.recognizer.Recognizer.allocate(Recognizer.java:182)
        at demo.sphinx.helloworld.HelloWorld.main(HelloWorld.java:54)

--- (Edited on 8/26/2008 2:09 am [GMT-0500] by Visitor) ---

Hi dasinen,

>now my problem is while using this created model in HelloWorld example it

>showing the following error help me on this

Sorry, I don't know much about Sphinx 4 (yet...)

Ken

--- (Edited on 9/1/2008 8:22 pm [GMT-0400] by kmaclean) ---

Hi,

Please tell me How to connect the output of this spinix speech recognition system to a speech enabled web browser. is this act as a plugin

--- (Edited on 10/12/2009 12:39 pm [GMT-0500] by rdus) ---

--- (Edited on 10/12/2009 1:14 pm [GMT-0500] by rdus) ---

Hi rdus,

>How to connect the output of this spinix speech recognition system to a

>speech enabled web browser.

You might want to look at how the WAMI toolkit was implemented.

Ken

--- (Edited on 10/12/2009 7:59 pm [GMT-0400] by kmaclean) ---