混合摄影：生成式视频重拍中的预可视化与幻觉风险管理
Hybrid Cinematography: Previsualizing and Managing Hallucination Risk in Generative Video Reshooting
Nhan Tran · Neal Wadhwa · Abe Davis · Stefan Stojanov
中文摘要
在电影拍摄中，镜头运动在拍摄时即已确定。生成式视频重拍允许后期修改，但可能需要幻觉未记录的内容，这一缺口有时在离开现场后才被发现。混合摄影（Hybrid Cinematography）工作流连接物理拍摄与生成式重拍，在创作者仍可行动时管理幻觉风险。利用可编辑的3D镜头计划和拍摄代理，预可视化实时评估幻觉风险。看到拍摄缺乏支持之处，创作者可迭代调整计划，探索平衡拍摄与生成的运动，拍摄引导性补拍，或明知故犯地接受幻觉。通过移动增强现实应用（用于现场规划、拍摄和审查）及离线视频处理管线展示该工作流。一项针对资深电影人的研究表明，风险预可视化如何影响镜头决策，并揭示创作意图与生成幻觉之间的张力。
关键要点
- 01解决生成式视频重拍中因未记录内容导致的幻觉风险滞后发现问题。
- 02提出混合摄影工作流，结合可编辑3D镜头计划与拍摄代理进行实时风险预可视化。
- 03提供移动AR应用用于现场规划与审查，以及离线管线处理现有视频。
- 04研究发现风险预可视化影响镜头决策，并暴露创作意图与生成幻觉间的冲突。
解读
尚无解读。
原始英文摘要
arXiv:2610.12455v2 Announce Type: replace-cross Abstract: On a film set, the camera move is committed during a take. Generative video reshooting lets filmmakers change it afterward, but may require hallucinating unrecorded content, a gap sometimes discovered only after leaving the set. We present Hybrid Cinematography, a workflow that bridges physical capture and generative reshooting to manage hallucination risk while filmmakers can still act on it. Using an editable 3D shot plan and a proxy of the take, our previsualization evaluates hallucination risk in real time. Seeing where the take lacks support, filmmakers can iteratively adjust the plan, explore moves that balance capture and generation, shoot guided pickups, or knowingly accept hallucination. We demonstrate the workflow through a mobile augmented reality application for on-set planning, capture, and review, and an offline pipeline for existing video. A study with experienced filmmakers reveals how previsualizing risk informs camera decisions and exposes tensions between creative intent and generative hallucination.