I currently use the legacy animation system to support lip synching, all other animations use mecanim. I'd rather use one system and I have to keep mecanim so I'd like to dump legacy.
The reason I don't use mecanim for viseme animation is that it is not scriptable (well, it's barely sort of scriptable) and I can't create blend trees dynamically during runtime. (At least I don't know how...)
Given that I use 9 visemes and a reference position, and they can all transition to each other, and they have to play at different speed based upon how long the viseme will be displayed (faster/slower talking/annunciation) - has anybody done something like this in mecanim?
I presume I could create a gigantic state machine, in its own layer, that held all the possible combinations (ye Gods and little boarlets...) but I'd still be stuck with the playback speed limitations would I not? (Can I add another animator just for running visemes and adjust that speed?)
Thanks :)
↧