Generating Found-Footage Fantasy Short Films with H3: Complex Prompt Test
Disastrous-Agency675 · reddit · 2026-08-06
The author shared an impressive "found footage" style fantasy short film generated using the video model H3.
The post details the complex prompt structure used:
- Camera Control: Specified that the first second fully references the input image, adopting a shaky handheld camera feel.
- Scene & Narrative: Set in a medieval fantasy world, an elf ranger speaks directly to the camera (viewer), while wolves wander off into the mist.
- Multi-Shot Scheduling: Cuts to a static shot at 7 seconds, simulating a camera placed on a log, advancing the story through lower-body framing and off-screen dialogue.
More from Multimodal
- WISRD Benchmark: Evaluating AI Problem-Solving in Pure Image Space — HirokatuKataoka · 2026-08-06
- AI Recreates 90s Seinfeld: Perfect Sitcom Vibe with Data Center & AI Jokes — FightingBlaze77 · 2026-08-06
- ComfyUI Workflow for Minimax H3 Turbo Lora Shared — scooglecops · 2026-08-06
- Minimax H3 Ref2V Quality Issues: Parameter Tuning and Discussion — ShengrenR · 2026-08-06
- FireRed-Image-Edit Fast Tops Hugging Face Trending Spaces — prithivMLmods · 2026-08-06
- OpenAI Reportedly Acquires Top Image Model Company Reve — cocktailpeanut · 2026-08-06