Seamless 1-Shot Lip-Sync with Minimax H3 via Per-Token Noise Masking
stonyleinchen · reddit · 2026-08-15
Release of a ComfyUI workflow update enabling seamless lip-sync using Per-Token Noise Masking on audio and video latents. This method outperforms guidance/reference-based approaches by enabling strong convergence from step 0 and running faster without latent expansion. It pins the music track to the latent space, protecting it from denoising to create strong conditioning, allowing perfect lip-sync even with the FL model. The repo includes custom nodes and example workflows.
More from Multimodal
- How meloform works: WASM pitch detection plus differentiable circuits learning 1,022 sine coefficients — henloitsjoyce · 2026-10-16
- Documentary filmmaker urges peers to watch 'You Can See Everything' — kyliebytes · 2026-10-16
- Redditor's first Cumfy/Seedance video nails a cinematic one-shot generation — Wave_N_Wind · 2026-10-16
- Kling 4 returns after 8 months: 10 hyper-realistic AI ads for under $5 — CurieuxExplorer · 2026-10-16
- Shanghai AI Lab unveils Intern Lumina U2, a diffusion LLM unifying image, video and 3D tasks — Shanghai-AI-Laboratory · 2026-10-16
- NAVER AI Lab proposes Speculative Forcing to speed up autoregressive video generation — naver-ai · 2026-10-16