Speech Recognition Engine Training

Hello
I'm using MS Speech Recognition Engine in my C# project ( System.Speech.Recognition default engine under Vista - .net framework 3.5. )
So, through SpeechRecognitionEngine class I'm able to recognize some letters and some simple words.
I found some problems in the letters case, the recognition is not as accurate as I though.
My program only needs to recognize one person voice at time So I need to know if its possible to train the Engine with that person
pre-recorded voice before start recognizing, so detection would be more precise.
Can you guys help me?
Thanks in advance
Alex

go here:

http://www.codeproject.com/KB/audio-...?display=Print

you will find some hint how to train it also you can download the source and see how it works:

in general you have two possibilities for training:

1- Mic training wizard
2- User training wizard