Creating A Limited Vocabulary Voice Recognition System using C++

[dj_redface]

Hello,

For a university project I am trying to write a program that will turn
vocal commands into MIDI events. For example if the user says A, a
note "A" will be played.

I am trying to write this program in C++ using the Microsoft Speech
SDK 5.1. I have written a library that drives the MIDI aspect, and am
able to write all the playback aspect. I have yet to find a way of
limiting the vocabulary to recognise specific commands. For example
letter names, flat/sharp/natural/ etc.

I would be grateful for help in limiting the window's vocabulary. Or
suggesting another way to create this program. I am happy to scrap
using the SDK as at the moment this is project is still in the
research stage.

Thank you.