Idea Exploration: Can PiD Upscaler Be Used as a Refiner Stage for Video Generation
bonesoftheancients · reddit · 2026-08-04
A creator explored a custom workflow idea for video post-processing in the community: due to the lack of specific PiD upscaling tools for certain videos (like H3), they plan to split the generated video into frames, pass them through an image model's (like zimage or flux) VAE encoder to extract latents, upscale directly in the latent space, and finally recombine them into a video. The author posted asking if this technical path is practically feasible.
More from Multimodal
- ComfyUI posts MiniMax H3 workflow examples for native text-to-video — Replikante · 2026-08-04
- A ComfyUI workflow pushes Wan 2.2 image-to-video out to roughly 45 seconds — embryo10 · 2026-08-04
- A tiny H3 demo detail: the weights shift during the conversation — Oatilis · 2026-08-04
- Reddit users compare Qwen3.5-VL, InternVL and Gemma 4 for uncensored image captioning — TekeshiX · 2026-08-04
- AI-made “Ancient China” video draws attention for its stylized visual design — xiaosun86 · 2026-08-04
- Qwen Image Edit crashes in ComfyUI when blur nodes are added first — UnorthodoxyMedia · 2026-08-04