Adobe’s Wonder turns images into a camera-controllable video world model
adobe · hf · 2026-07-29
Adobe Research and Johns Hopkins present Wonder, a general-purpose video world model for real-time, camera-controllable exploration. Starting from an image or a conditional video, it builds a playable world where users can move the camera, revisit explored regions, and discover unseen areas over long horizons.
Key ingredients include:
- Dense coordinate-field camera conditioning to turn camera motion into visual evidence.
- Sparse-attention memory for fast retrieval over long generation contexts.
- Distillation fixes that help the student model obey control signals while keeping diversity and long-term memory.
The system can synthesize minute-scale videos at 16 FPS and also supports video-conditioned re-shooting of existing dynamic scenes in real time.
Related event: Adobe and JHU Unveil Real-Time 3D World Model 'Wonder'(3 posts)→
More from Multimodal
- Seedance 2 Video Model Test: Life in a Western Town — Necessary-Use-3820 · 2026-07-30
- Creator Debuts First Episode of AI-Generated 'Odyssey' Short Film — DavidmComfort · 2026-07-30
- Wonder: Real-Time Interactive 3D World Model from Adobe and Johns Hopkins University — yuyinzhou_cs · 2026-07-30
- Fan-Made Music Video for Skrillex's 'Kyoto' Created Entirely with AI — Procyon87 · 2026-07-30
- nano banana pro in Action: Generating Cinematic 3x3 Storyboards — techhalla · 2026-07-30
- Creator Recreates a True Story Using AI Video Generation — Daniel_Ai_VFX · 2026-07-30