Wan Video Generation Struggles with Hand Artifacts, Author Seeks Help
alexmmgjkkl · reddit · 2026-07-07
A user reported that when using Wan (Tongyi Wanxiang) for image/video generation, hand artifacts like blurriness and detail loss persistently appear regardless of the model or workflow used (animate, scail, steady dancer, phantom). The issue is especially noticeable every other frame, and adjusting steps and CFG yields no results.
Post-processing with Klein achieves near-animation frame quality, but since Klein lacks awareness of animation sequences, processing image sequences introduces new issues. The author wonders how those clean hand effects on YouTube are achieved, suspecting limitations related to fp8 models, Kijai's ComfyUI wrapper, or the subject matter.
More from Multimodal
- Dense instance segmentation via Muse Wonder + SAM 3.1 on Meta's Model API — nikhilaravi · 2026-10-26
- LichtFeld Studio v0.5.4.3 ships community-built SteadySplats, POPSpa and Voxelization — janusch_patas · 2026-10-26
- One prompt to a 2-3 min narrated short: hands-on test of Antigravity video generation — AI_Andrew · 2026-10-26
- Creator makes 10,564-frame animated short entirely with Gemini toolchain — AI_Andrew · 2026-10-26
- FLUX.3 Klein step-distilled to 4-8 steps with CFG baked in — Fit-Structure-6489 · 2026-10-26
- ostris pre-training "Primal VAE": 1-channel latents match modern 8x VAEs via patching — ostrisai · 2026-10-26