Hello :-)!

I need very much to create simple application which uses speech recognition. In order to do it, I need to create acoustic model for my application (which is not supported here in VoxForge). My vocabulary is really little and it contains about fifteen words.

I know I need to follow this tutorial http://www.speech.cs.cmu.edu/sphinxman/scriptman1.html but I've got some difficulties with this. This is why I decided to download Germany acoustic model and try to do my model analogically to Germany model.

I am aware of the fact that I ask for much but I would be really greatful if somebody experienced can guide me through the process of training acoustic model. I see you have spent really much time on training Germany model so I am sure you would be able to help a newbie a little bit :-). In order to avoid redundant posts, please visit this topic: http://www.voxforge.org/home/forums/message-boards/acoustic-model-discussions/creating-new-model-with-the-use-of-germany-voxforge-model#FjNVxta-IKzo76sePraN1g .

Greetings :-)!

>This is why I decided to download Germany acoustic model and try to

>do my model analogically to Germany model.

It is best to stick with your own language.

Ken

Thanks for your answer!

I still wait for the answer on the forum topic mentioned above. I follow the other tutorial (http://www.speech.cs.cmu.edu/sphinx/tutorial.html#train), I used script to copy an4 and create new acoustic model. The tutorial says that I need to change dic, filler and phone files. That's understandable and I'm going to do it. But other very important thing is not only to write list of files but also to record them. So I enter wav\an4_clstk directory and I see files .sph, not .wav. It is not explained in that tutorial why I've got sph files and why there are so many directories (like fash, fbbh etc.). How to create recordings (where to copy them? can it be done with default Windows recorder in Accessories?).

Greetings :-)!

record in wav and convert to sphere format using SoX