Notably, at the time of release, Microsoft said the MAI-Voice-1 is a speech generation model that natively generates expressive and natural-sounding voice. It can generate a full minute of audio in ...