Home Artists Posts Import Register

Content

Believe it or not there still isn't a good quality solution for AI speech synthesis. The few that exist require hours of audio for a good output.

Currently the best options I've found so far are ElevenLabs, Play.ht, TortoiseTTS, and Bark. ElevenLabs won't have professional voice cloning until later this year and will cost 22$/month. Play.ht is a good free one but requires really good training audio which will require quite some time to put together. My current data sets are too small for good results and must be high quality audio clips without too much emotion. Bark is a new one with great potential but currently the only problem it faces is the output is a bit robotic. 

Aside from that when it comes to the actual development of the game I've been planning on starting to work on it again in the summer. I should hopefully have extra time.

Comments

No comments found for this post.