Faceswap Forum

Is there a way to change/control the lip movements of a character while keeping the original video setup?
I can't seem to find any workflow for this task

This is not a supported usecase for faceswap. Faceswap focuses only on face-replacement. I'm not aware of any software designed for lip re-enactment.

As far as i understood, no way on doing it directly with faceswap or similars... I'm looking for the same effect and have found many articles on programs doing just that, but not avaible on open source (or for sale either)...

I don't have all the sources found but here is one (i know there is at least two more, but can't find them:
http://grail.cs.washington.edu/projects/AudioToObama/

Otherwise, it seems like you have to go back and forth with the process. Swap the target's face on you first, make your speech, then put your swaped face back on the target. It sounds like a toutchy process since you have to match global cordination of the head and use some fine compositing.

I unfortunatly am on a mac with dumbass amd graphic cards and am just not smart enough (yet) to have made the google collab method work. I also tried out the "first order motion model" (https://colab.research.google.com/githu ... demo.ipynb) method, but the size of the result is simply too poor (256p) for my use, and i don't believe in the upscaling process to be abble to upscale it to a 720p at least.

If anyone else has a few tricks or tips on metodology (VFXChris Ume here? i think he uses deepfacelab... But go check his work on YT).

M.