Today I heard speech recognition expert Dan Klein describe this ATT Research project as something close to the ‘top of the line’ in text to speech synthesis.

I have no real understanding of the challenges of natural-sounding voice synthesis from text, but it’s fascinating to hear this and think of it as the best (or close to the best) that we can do these days. It’s certainly a great leap forward from the standard text to speech synthesizers that are included with Windows, for example. Have fun playing around with it…