search

zhangzc21 / DynTet

178 stars 21 forks source link

3DMM audio coefficients #7

Open manalHassan opened 3 months ago

manalHassan commented 3 months ago

Hello! I've been exploring the DynTet repository, and I came across the note in the README that mentions using SadTalker to generate the 3DMM coefficients given audio. The note states, "Note: Given an audio, you can try SadTalker to generate the 3DMM coefficients."

I followed the instructions, installed SadTalker, and ran it as directed. However, I did not see any npy file being produced. I'm particularly interested in obtaining another audio npy file.

Could you please provide more details on how I can obtain this file? Is there a specific audio input or additional steps required to generate the npy file using SadTalker?

Thank you for your assistance! Best regards, Manal

zhangzc21 commented 2 months ago

Hi, Manal, you can directly generate "FILENAME.mat" from "WAVENAME.wav" file by sadltalker, then try

python infer.py --config configs/obama.json --drive_3dmm FILENAME.mat --audio WAVENAME.wav

Actually, the npy file we provided was converted from the mat file.

hereTac commented 2 months ago

You could reference the ISSUE #13 about the .npy or .mat . @manalHassan