A 6-site workflow for making AI videos look cinematic, from shot lookup to lighting
DavidmComfort · x · 2026-10-18
A Japanese creator shared a step-by-step workflow for making AI-generated video look cinematic using six sites in order: SHOT.CAFE (searchable database of 24,321 film stills with 109,941 tags across 187 films) for reference shots, browsing classic compositions and colors, learning cinematography technique names, translating techniques into prompts, converting camera movement into prompt language, and finally refining lighting and editing. A practical prompt-engineering pipeline grounded in real film language.
More from Multimodal
- AI-generated news anchor fools viewers in Paris broadcast clip — djcows · 2026-10-18
- MOSS-Image 5B open-weights model stuns with accurate text layouts, ~15s per local image — alexcovo_eth · 2026-10-18
- SteadySplats runs in-browser at 240 fps with near-free training overhead — bilawalsidhu · 2026-10-18
- Testing Gemini 4 Argon on court mapping: solid but position errors remain — Roger_M_Taylor · 2026-10-18
- AI-made animated kids series 'THE SUMMER DAY' with original song — Warblewick · 2026-10-18
- Turn Claude into an AI video production studio with six small jobs — Roger_M_Taylor · 2026-10-18