lip-sync to audio
#32
by timtianyang - opened
Hi, I've trying to adapt the workflow here to use audio as input and prompt the character to lip-sync to it. Instead of using empty audio latent, I pass the audio latent to the latent concat node.
It worked somewhat, but it appears the generate audio is missing certain syllables and the character therefore doesn't lip sync correctly like 10% of the time. Do you have any suggestions in making this into an IA2V workflow? Or am I missing some assumptions on the limitation here.
Thanks!