updream's new previs stage turns reference images into 3D white models for AI video shot control
APPSO · wechat · 2026-08-17
APPSO tested updream's new previs stage: no modeling skills needed—upload up to 3 scene reference images and get a 3D white-model scene in 4-7 minutes, where creators place characters, set camera positions, trajectories and keyframes, converting prompt-dependent shot design into visual previsualization. It's described as a Blender for AI video creators.
Tests across mech hangar, cosmic hub, subway pass-by, rooftop standoff and wuxia duel scenarios showed that adding white-model reference made camera moves and timing far more stable, with fewer deviations in framing and reveal moments. Coarse white models handle blocking and camera routes, while action details remain with prompts and models like Seedance2.5—each control channel managing one type of information. For complex shots costing tens of yuan per generation, previs also cuts re-roll costs.
The article draws an analogy to Kodak's Brownie camera, arguing such tools democratize cinematographic language. updream also launched a white-model creation challenge with up to 1 million credits (worth ¥10,000) per top entry.
Related event: updream Launches Previs Stage: 3D White Models Tame AI Video Camera Control(4 posts)→
More from Multimodal
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- Open-source anatomy-reveal skill turns Claude into an exploded-view motion design studio — msharmas · 2026-10-16
- Flux 3 Klein vs Flux 2 Klein hands-on: the new base model 'barely usable' for T2I — Both-Rub5248 · 2026-10-16
- USTC paper uses speculative sampling to cut latency in interactive video generation — ustc · 2026-10-16
- 800,000 Gaussian splats run in WebGPU; Casado sees near-infinite scaling — generativist · 2026-10-16
- Nano Banana 2.1 image generation now live in Perplexity Computer for all users — ChrisUniverse · 2026-10-16