I noticed while watching the Documentary about how the Simpsons used CH to animate the live section with Homer, how Dan Castellaneta read something called a Toothsome Meme to train CH and help the software recognize his voice.
Is that a step that the users should do? If so, how should we proceed exactly?
We don't currently have this capability on the user side, but it would be cool one day to read a poem like this and have the lip sync more tuned to your own voice. I'll add you to the list of supporters for this as a potential future feature!