21 ms·
As noted in other comments, all the samples still sound very robotic, so this is probably "just" a method to tune the parameters of an existing voice synthesize
by kristaps 9y ago
As noted in other comments, all the samples still sound very robotic, so this is probably "just" a method to tune the parameters of an existing voice synthesizer to mimic a real persons voice as much as it allows.
- TillE 9y agoThat's exactly what it sounds like. The same old mediocre TTS with voices modified to mimic specific well-known voices. It's impressive for what it is, but a lot of people here seem way too excited. This isn't any kind of breakthrough, and only the shortest hand-picked snippet would fool anyone.