AI filmmaking's hardest problem: Reddit practitioner seeks cutting edge for multi-shot character and spatial consistency
Due-Air-5626 · reddit · 2026-11-04
Reddit user Due-Air-5626 posted a detailed thread on AI filmmaking's core challenge: not generating pretty single shots, but maintaining character identity, spatial consistency, precise blocking, camera control and shot-to-shot continuity across a scene. They've tried a Blender → AI video pipeline (Mixamo characters, animated cameras, depth maps, pose conditioning, reference images, MiniMax H3 via ComfyUI) with promising single-shot results but unreliable spatial consistency, and now question whether forcing general video models to respect 3D blocking is fundamentally wrong. They ask about 3D-as-ground-truth pipelines, persistent identity approaches (LoRAs, identity adapters), Gaussian splatting/world models, node-graph engineering, and what actually works in production.
Related event: Multi-Shot Consistency Remains the Hardest Problem in AI Filmmaking(2 posts)→
More from Multimodal
- Dev uses Gemini 4 + Antigravity to auto-document a week of building into a BTS video — alexanderchen · 2026-11-05
- Ex-Google AR founder launches Atmee v2: real-time avatar model claims 50x speed, 100x cheaper — eptwts · 2026-11-05
- Creator shares Frozen + reverse Seedance video prompt combo — techhalla · 2026-11-05
- IV Studio: open-source local film studio on ComfyUI + MiniMax H3, idea to short film — itsVariance · 2026-11-05
- Running open source image/video models without a GPU: CozyGPU bills raw GPU time, cheaper than marked-up platforms — MissMoneyBalls · 2026-11-05
- From stills to a playable game-like world: a Weave creative workflow (plus discount code) — gizakdag · 2026-11-05