AI
D\'ecoupage: Directorial Audio-Visual Control for Long Multi-Shot Video Generation
Publication: cs.AI updates on arXiv.org
Author: Yifan Wang, Hao Ouyang, Qiuyu Wang, Yihao Meng, Wen Wang, Zeyue Tian, Yichong Lu, Hanlin Wang, Qingyan Bai, Zichen Liu, Yujun Shen, Qifeng Chen
Source published: 17/11/2026, 05:00:00 UTC
Briefing created: 17/11/2026, 10:53:11 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 17/11/2026, 08:33:24 UTC.
Briefings
DECOUPAGE is a screenplay-driven framework for long-form video generation that uncouples visual and acoustic tracks to coordinate independent timing. It builds on a joint audio-video diffusion model with dual-schedule temporal cross-attention and a directorial Lip-sync mechanism that selectively decouples speech conditioning from visual facial dynamics during off-screen dialogue. On multi-shot conversational benchmarks, the system achieves precise control over shot transitions, audio timing, and visibility-aware lip synchronization, directing 15- to 60-second generation with multimodal coherence.
Key points
- DECOUPAGE is a screenplay-driven framework for long-form video generation that uncouples visual and acoustic tracks to coordinate independent timing.
- The system achieves precise control over shot transitions, audio timing, and visibility-aware lip synchronization on multi-shot conversational benchmarks.
Context
Existing audio-visual generative frameworks lack explicit mechanisms to coordinate independent visual and acoustic timing over extended horizons, often hallucinating speech during intended silences. DECOUPAGE addresses this by building on a joint audio-video diffusion model with dual-schedule temporal cross-attention and a directorial Lip-sync mechanism.
Limitations
The evidence is from an arXiv preprint (arXiv:2611.18866v1) and does not include experimental results, user feedback, or real-world deployment data. The study uses multi-shot conversational benchmarks rather than specific city or country locations, limiting geographic scope.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 17/11/2026, 10:53:11 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeigpwvu4fmxlxitz3xkwr26rs7tpperjmpydfqnrv6r3ih24gxcjdi