MiniMax H3 music video clip showcases character consistency and lip-sync
The clip was made with a full-reference workflow: character reference sheet, location reference, and an isolated vocal stem for lip-sync, with no music in the audio input; everything else was generated by the model. The creator says consistency and lip-sync held up across shots.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Reddit user shares Minimax H3 character/object V2V swapping template
- Reddit user shares AI video made with Krea H3 and ComfyUI
- Angelo: Video Edition coming soon for ComfyUI
- DeepZero automates Windows kernel vulnerability research with AI agents
- Compact Magazine essay argues AI will destroy Americans' way of life