Tri-Prompting: Video Diffusion with Unified Control over Scene, Subject, and Motion

TL;DR

Tri-Prompting introduces a unified video diffusion framework that enables precise control over scene, subject, and motion, enhancing multi-view consistency and identity preservation for versatile content creation.

Contribution

It presents a novel unified architecture and training paradigm that integrates scene, subject, and motion control in video diffusion models, supporting complex, multi-faceted video editing tasks.

Findings

Outperforms specialized baselines in multi-view subject identity preservation

Achieves superior 3D consistency and motion accuracy

Enables novel workflows like 3D-aware subject insertion

Abstract

Recent video diffusion models have made remarkable strides in visual quality, yet precise, fine-grained control remains a key bottleneck that limits practical customizability for content creation. For AI video creators, three forms of control are crucial: (i) scene composition, (ii) multi-view consistent subject customization, and (iii) camera-pose or object-motion adjustment. Existing methods typically handle these dimensions in isolation, with limited support for multi-view subject synthesis and identity preservation under arbitrary pose changes. This lack of a unified architecture makes it difficult to support versatile, jointly controllable video. We introduce Tri-Prompting, a unified framework and two-stage training paradigm that integrates scene composition, multi-view subject consistency, and motion control. Our approach leverages a dual-condition motion module driven by 3D tracking points for background scenes and downsampled RGB cues for foreground subjects. To ensure a balance between controllability and visual realism, we further propose an inference ControlNet scale schedule. Tri-Prompting supports novel workflows, including 3D-aware subject insertion into any scenes and manipulation of existing subjects in an image. Experimental results demonstrate that Tri-Prompting significantly outperforms specialized baselines such as Phantom and DaS in multi-view subject identity, 3D consistency, and motion accuracy.