search

yunik1004 / SAiD

SAiD: Blendshape-based Audio-Driven Speech Animation with Diffusion
https://yunik1004.github.io/SAiD/
Apache License 2.0
74 stars 14 forks source link

Face motion keys not matching problem #7

Open P2Oileen opened 4 months ago

P2Oileen commented 4 months ago

Nice Work! I have noticed that the blendshape format works for anime characters very well. I am trying to reproduce the result in your websitehttps://yunik1004.github.io/SAiD/, especially for "Visualization on different blendshape facial models" part. I downloaded the 3D model, but find the keys is not the same as those in the csv file. For example: image For mouth motions, the 3D model have keys above. However, in csv files, those keys are different:

mouthClose | mouthFunnel | mouthPucker | mouthLeft | mouthRight | mouthSmileLeft | mouthSmileRight | mouthFrownLeft | mouthFrownRight | mouthDimpleLeft | mouthDimpleRight | mouthStretchLeft | mouthStretchRight | mouthRollLower | mouthRollUpper | mouthShrugLower | mouthShrugUpper | mouthPressLeft | mouthPressRight | mouthLowerDownLeft | mouthLowerDownRight | mouthUpperUpLeft | mouthUpperUpRight

How can I match those keys to perform correct video results like you did in the video on your website? I really appreciate your help. Thank you again for open sourcing this facinating work.