Viseme format - Githubissues

liukuangxiangzi / audio2viseme

The code generate phoneme from audio features.

23 stars 3 forks source link

Viseme format #1

Open samsgates opened 3 years ago

samsgates commented 3 years ago

Is this generate lipsync viseme ID details? like this identification : https://www.researchgate.net/figure/Example-visemes-for-phoneme-classes-given-in-Table-1-21-Phone-Based-Acoustic-Units_fig1_4273506

liukuangxiangzi commented 3 years ago

Is this generate lipsync viseme ID details? like this identification : https://www.researchgate.net/figure/Example-visemes-for-phoneme-classes-given-in-Table-1-21-Phone-Based-Acoustic-Units_fig1_4273506

I am using 'The Disney twelve phoneme-to-viseme map' from this paper: https://arxiv.org/pdf/1710.01288.pdf

samsgates commented 3 years ago

Thanks for your response.

Do you have viseme sample model with 12 blend-shapes? or any sample to create blend-shapes in our existing model. I didn't find sample lip shape for each viseme.
How to calculate the -d and -c value (Delay frames, number of context frames), is that mandatory?
Do you have any idea how to apply this viseme in to unity 3D model lipsync