Audio 2 face blendshape generation error

when i finshed mesh post swap, try to export my blendshape, the audio 2 face has this error:

[Error] [omni.graph.core.plugin] /World/LazyGraph/deltaBlend_space_ReadPrimAttribute: inputs:name is empty

any one can help?

Hi there,

Can you please send your latest Audio2Face log file which can be found here: C:\Users\<USERNAME>\.nvidia-omniverse\logs\Kit\Audio2Face

Also, looking at your scene might help find the issue. Is it possible to share your file too?

kit_20230403_004416.log (1.4 MB)

I downloaded the log file from audio2face console.
I didn’t shut the app down, hope you guys can work it out soon.
I also have another question about the mesh fitting, why it has to separate the mesh to fit the shape key?
For instance, my character’s eyelid and the head didn’t make from the same mesh, I try to join the eyelid to the head into 1 mesh before load into the audio2face.
It sent me warning( I sure you guys can found out from the log file?)

Though I did it again without the eyelid, audio2face still fail to export blendshape.

Hope this can be solved too.

Thank you very much.

111_usdSkel.usd (8.5 MB)

kit_20230403_004416.log (1.4 MB)

I created a new setup from scratch using the Rabbit head and was able to export blendShapes successfully. I’m not quite sure at which stage you’re receiving this error.

Can you share a step-by-step guide which results in this error?

Resulting UsdSkel BlendShape
bunny_bs_usdSkel.usd (5.7 MB)

Hi, thanks for your help.
Here are my steps.
1.creat usdc file in omniverse enbeded blender.
2.import the usdc file into audio2face.
3.set up meshes
4.mesh fitting wrap
6.export my file
7.go back to blender, transfer shape data, it failed.
8.check blender log, it said missing sth.
9.go back to audio2face,check the console,a bunch waring too.

Please help, thanks.
log file.rar (11.9 MB)

Some steps missing between step 5 and 6. After post-wrap, you need to assign an A2F pipeline to grey Mark head, then generate BlendShapes. This will only export BlendShapes, not the weights, so you have to use A2F Data Convertion tab to generate weights for your newly created BlendShape and export the weights to USD or JSON after that.

Please take a look at our BlendShape generation and export tutorials.
Audio2Face with Blender | Part 1: Generating Facial Shape Keys - YouTube
Audio2Face with Blender | Part 2: Loading AI-Generated Lip Sync Clips - YouTube

1 Like

Thank you very much

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.