it just significantly reduces motion. i imagine if I ran through 2 samplers I might be able to get some back but its tricky and the render time would be quite long
use cumshot loras, write a prompt about a guy walking in out of frame and cumming on face
insert picture of a photograph, man will walk in, cum on the face of the person in the photo and the person will react and the cum will get on their face
amazing visual effect would be an absolute bitch to pull off without AI
i have just now discovered that i can just tell wan to generate more than 81 frames and it just does it. 10 second continuous. i want to see how far it goes before OOM lol
the one thing i noticed is that the movement is more subdued, if I increased the steps maybe it would be better but then render times would be kind of high