46 blendshape standards of audio2face

AI & Data Science Digital Humans (closed) Audio2Face (closed)
1
  1. Is there ready-made code to convert 46 blendshapes of audio2face into 52 blendshapes suitable for arkit？
  2. Can you provide schematic pictures corresponding to 46 blendshape to show the corresponding effect of controlled facial expression？
2

Hello @fantastic_wn! I will ask the team for help in answering your questions. I will post back when I have more information!

Also, check out this video here: BlendShape Generation in Omniverse Audio2Face - YouTube at around 2:23 in the video, you can see the 46 blendshapes that were generated.

And, if you haven’t already, you can check out our Audio2Face Documentation here: Audio2Face Overview — Omniverse Audio2Face documentation

3

Hi @fantastic_wn,

  1. We don’t have the specific code at the moment, but you can check this tutorial and modified the given script? Very similar process of mapping 46 blendshape weights to different controls.

  1. You mean you want to review the individual shapes in the blendshape model? If so, you will need to check the shape from .usda file (ascii format) or by feeding different blendShapeWeights to the skelAnimation node. In the upcoming release, we will make it easy by providing a button that shows all the shapes in timeline.