Hi! Yea, i'm also found exactly same wf as effective and with an expected result (screenshot > convert to stylized image > realism look) . I used qwen image edit 2511 and wan2.2 for video gen. Also 3 shots created with ltx-2, but i don't like the quality of that model and long prompting to get some acceptable output. But i like the possibility to make audio and lipsync with just image and prompt and very fast. The last shot with Zombie as example.