Fighting MiniMax H3 in ComfyUI: static camera, clean background, full-body room
Kassiber · reddit · 2026-09-28
The author uses MiniMax H3 in ComfyUI to animate full-body illustrated characters for a 2D game but hits three issues: the camera zooms despite static-camera prompts, tight framing cuts off hands/feet, and with more empty space H3 invents backgrounds (once adding a huge duplicate character behind). Setup: FL2VA INT8, Turbo 8-step v1.0, 124 frames, RTX 5070 Ti 16GB, tested 1280×2240 and padded 1792×1792 canvases. They ask for a repeatable workflow: tested prompts, background choices, framing control without losing facial detail, and smooth transitions between consecutive action clips.
More from Multimodal
- How to Recreate the Viral AI Halloween Door Video — motionbynick · 2026-09-28
- Style Atlas: a gallery of 1,337 art styles with copyable prompts for AI image generation — mhmazur · 2026-09-28
- Qwen Image character sheet workflow: one reference image to a full design sheet — LeebLaab · 2026-09-28
- Moiré Pattern Interference prompt: overlapping fine lines that fake motion and depth — LudovicCreator · 2026-09-28
- A production-line workflow for pro motion design with Opus 5.5: storyboards first, render last — sven_ai · 2026-09-28
- RTX 3060 shootout of five image models: Qwen Image 2.1 trails ERNIE and Krea 2 — SirMick · 2026-09-28