I'm currently working on the next episode of my AI series about humanity's future in the cosmos (https://www.youtube.com/watch?v=X1NIjn_kJq8). And this time I wanted to have a talking character, a presenter who becomes the protagonist of the short film.
You may remember that in one of my older videos (https://youtu.be/nTLuH-uRaZ8?si=7ukgU_3pztAFC2ua) I created a pretty complex workflow for this. I used "Wav2Lip", which produces really good LipSync for free, but unfortunately with a very poor image quality.
I was quite happy with the results, but creating these shots took a lot of time because of the many steps involved. That's why I simplified and automated the with ComfyUI.
Download the Workflow here! And don't forget to play around with the strength (codeformer fidelity) and model (Codeformer or GFPGAN)!
Moon Yun Pellerin
2024-02-26 01:23:02 +0000 UTCMoon Yun Pellerin
2024-02-26 01:19:59 +0000 UTC