2 ms·
Google and Baidu have only demonstrated single speaker TTS. Lyrebird's the first to demonstrate being able to generate arbitrary voices. Since this came out of
by backpropaganda 9y ago
Google and Baidu have only demonstrated single speaker TTS. Lyrebird's the first to demonstrate being able to generate arbitrary voices. Since this came out of a research lab, I would guess that the quality would only improve if they are given more compute and data.
- jtbayly 9y agoThe Google one is able to generate arbitrary voices based on the recordings used for training. So much so that they made it generate piano music.