voice recognition / dialect sensitive

The terrible twos's icon

Hi, im looking for any info or earlier post on voice recognition and dialect sensitive capabilities.

Basically i'd like to have a conversation with a sampler, having pre-recorded audio answers and statements in response to multiple groups of keywords, each with various possibilities (if/and/then...duh). I suppose ideally it would run similarly to an old chess program or dumber.

What are the voice recognition capabilities with max/msp? Are there any capabilities for dialect sensitive functions?

Im just digging up info to see where it goes.

Anyone interested in u.s. shares shoot me an email: strataradan@yahoo.com

Thanks - rob