4 ms·
All for suggestions! We've tried white noise as well, but it only works on plain talking samples (not music, for example). My guess is that the most robust solu
by andrew-w 2y ago
All for suggestions! We've tried white noise as well, but it only works on plain talking samples (not music, for example). My guess is that the most robust solution will come from updating how it's trained.
- bobbylarrybobby 2y agoWhat if you train it to hold the last frame on silence (or quiet noise)?
- andrew-w 2y agoWe've talked about doing something like that. Feels like it should work in theory.
- jazzyjackson 2y agoOr noise corresponding with a closed mouth Hmmmmmmmm Ohmmmmmmm