Generative Photographic Control for Scene-Consistent Video Cinematic Editing

2025-11-16

AI Summary

Existing generative video models lack fine-grained, disentangled control over professional cinematographic parameters—such as depth of field and shutter speed—limiting cinematic storytelling capability and spatiotemporal scene coherence. To address this, we propose CineCtrl, the first video editing framework enabling cinematography-grade parameter control. Our method introduces a disentangled cross-attention mechanism that explicitly separates camera motion modeling from photographic effect modeling, ensuring parameter controllability and inter-frame temporal consistency. Furthermore, we construct a large-scale video dataset with precise, physics-informed and real-world cinematographic parameter annotations. Experiments demonstrate that CineCtrl accurately honors user-specified multi-dimensional cinematographic parameters, generating high-fidelity, temporally coherent cinematic videos. It significantly outperforms prior methods in depth-of-field and exposure control tasks, establishing new state-of-the-art performance in parameter-aware video editing.

Technology Category

• Computer Vision: Computational Photography, Image & Video Synthesis
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Cinematic storytelling is profoundly shaped by the artful manipulation of photographic elements such as depth of field and exposure. These effects are crucial in conveying mood and creating aesthetic appeal. However, controlling these effects in generative video models remains highly challenging, as most existing methods are restricted to camera motion control. In this paper, we propose CineCtrl, the first video cinematic editing framework that provides fine control over professional camera parameters (e.g., bokeh, shutter speed). We introduce a decoupled cross-attention mechanism to disentangle camera motion from photographic inputs, allowing fine-grained, independent control without compromising scene consistency. To overcome the shortage of training data, we develop a comprehensive data generation strategy that leverages simulated photographic effects with a dedicated real-world collection pipeline, enabling the construction of a large-scale dataset for robust model training. Extensive experiments demonstrate that our model generates high-fidelity videos with precisely controlled, user-specified photographic camera effects.

Problem

• Controls photographic effects like depth of field in generative video models
• Disentangles camera motion from photographic inputs for independent control
• Addresses training data scarcity through simulated and real-world collection

Innovation

• Decoupled cross-attention mechanism for motion disentanglement
• Comprehensive data generation with simulated effects
• Fine-grained control over professional camera parameters