V10 the more I increase the passages the uglier and more distorted it becomes?
V10 the more I increase the passages the uglier and more distorted it becomes?
That is something you can't change. I try to enhance it a bit in a wf I'm working on, but WAN is made for 5 second videos. If you have multiple images of your character you could go with a start frame/end frame workflow. The end frame would be the start frame of the next batch. That would keep your charater way more consistent trough the hole video.
Thank you
Text to video joy
Try this author's workflow: https://civitai.com/models/1973861/wan22-all-in-one-mega-nsfw-infinite-video-generationvace?modelVersionId=2236647 . I use it to create a one-minute video script with virtually no loss of color consistency or quality. Watch the author's video for a better understanding. You can create a whole sequence of events from cues, stitched together into one long video.
Thank you