REMAINDER: a sci-fi short made with LTX and MiniMax H3 in ComfyUI, using restrained camera language
magratheya_64 · reddit · 2026-09-15
The author shares the full workflow behind REMAINDER, an AI sci-fi short built in ComfyUI, testing whether limited camera language — frontal, profile and top-down compositions, large architectural scale, flat color fields, minimal character motion — can hold visual consistency.
Workflow: still-image style development, LTX I2V for restrained environmental/character movement, MiniMax H3 for shots needing stronger motion or performance, separately generated voices, then edit/sound/mix in post. Deliberately avoiding photorealism in favor of flatter surfaces and graphic color separation. Story: a maintenance worker restores an old system in an abandoned civic facility — something answers.
More from Multimodal
- Dev cracks image conditioning on a DIY video model trained on one hour of footage — pixlpa · 2026-09-15
- Meta's Muse Voice Transcribe Goes Live in LiveKit Agents with 20+ Speaker Diarization — armand_ruiz · 2026-09-15
- 24-Second AI 'Fire God Origin' Short: One Continuous Shot, Speed Ramps Only — The_boneguy · 2026-09-15
- Creator puts himself on Survivor using Seedance 2.5 video generation — chrisfirst · 2026-09-15
- Image Tokenizers Define the Visual Language of Unified Multimodal Models — peterxichen · 2026-09-15
- Fiberwise Optimal Transport Schedules Cut Flow Matching FID by 38.6% — MaxUnfried · 2026-09-15