Sapi can also be made more responsive if you read the audio data into memory and then trim silence from the beginning of the speech. Many voices do not distinguish between a phrase break and the beginning of a new piece of text, so they actually insert silence in the beginning. If you trim that, it becomes very responsive. I wrote an implementation a few weeks ago that did this, and it makes a world of difference.
-- Audiogames-reflector mailing list Audiogamesemail@example.com https://sabahattin-gucukoglu.com/cgi-bin/mailman/listinfo/audiogames-reflector