hi all

for a project i might involved in we need to do some voice recognition. the setup is as follows : there is a virtual avatar that is able to perform certain movements. It should move its moth acording to the voice from the computer's sound input. So we do not need to recognise certain words or sentences, we just need to match the human voice with the movements of the virtual animation. This information would be sent to the animation software by OSC or some other method, this is not important right now.

I am not sure if this can be done with PD or some other software. we have an old speech detection system made in Java but it only works for wovels so it is not so realistic. Has any of you done something similar?

thanks!

enrike