MiniMax H3 pure text-to-video test with a full shot-by-shot prompt
BitterAd8431 · reddit · 2026-08-28
A frequent MiniMax H3 user tried prompt-only text-to-video (usually they use reference images), with an integrated LLM helping expand the prompt, producing a comedic manga-style clip of a white-haired cat-eared girl triggering a chain of kitchen fires.
The full prompt is shared, with a reusable structure:
- Shot-by-shot description covering character appearance, actions, scene changes and plot beats;
- Second-precise timeline control: e.g., pure nonverbal action from 0–2.88s, dialogue from 2.88–9.88s, reactions/ambience filling 9.88–14.38s;
- Audio-visual constraints: no human voices, breathing, or mouth movement outside the tagged dialogue window, with continuous ambience and synced SFX.
More from Multimodal
- The math: 5 rerolls at 480p plus a 4K upscale costs $5 vs $9 at 720p — HeyAmit_ · 2026-10-15
- Dreamina workflow: generate at 480p, upscale to 4K, cut video costs 44% — HeyAmit_ · 2026-10-15
- One-prompt GPT-6 Astra writes two Palestrina-style motets in LilyPond — deanwball · 2026-10-15
- DiffSynth-Studio Adds OPD Workflow: Few-Step Diffusion Training in ~2 GPU-Hours — bdsqlsz · 2026-10-15
- AI film and TV hit the market: Prime Video picks up AI spin-off, pilots screen at Mipcom — lmoroney · 2026-10-15
- Ramayana film taps DNEG's massive VFX and AI-assisted production for global release — CurieuxExplorer · 2026-10-15