MMControl: Unified Multi-Modal Control for Joint Audio-Video Generation

TL;DR

MMControl introduces a unified framework for multi-modal control in joint audio-video generation, enabling fine-grained, composable control over multiple conditions including visual and acoustic signals.

Contribution

It presents a dual-stream conditional injection mechanism and modality-specific guidance scaling for enhanced multi-modal controllability in diffusion-based models.

Findings

Achieves identity and timbre consistency in generated videos and audio.

Enables independent, dynamic adjustment of control signals during inference.

Demonstrates fine-grained, composable control over multiple generation aspects.

Abstract

Recent advances in Diffusion Transformers (DiTs) have enabled high-quality joint audio-video generation, producing videos with synchronized audio within a single model. However, existing controllable generation frameworks are typically restricted to video-only control. This restricts comprehensive controllability and often leads to suboptimal cross-modal alignment. To bridge this gap, we present MMControl, which enables users to perform Multi-Modal Control in joint audio-video generation. MMControl introduces a dual-stream conditional injection mechanism. It incorporates both visual and acoustic control signals, including reference images, reference audio, depth maps, and pose sequences, into a joint generation process. These conditions are injected through bypass branches into a joint audio-video Diffusion Transformer, enabling the model to simultaneously generate identity-consistent video and timbre-consistent audio under structural constraints. Furthermore, we introduce modality-specific guidance scaling, which allows users to independently and dynamically adjust the influence strength of each visual and acoustic condition at inference time. Extensive experiments demonstrate that MMControl achieves fine-grained, composable control over character identity, voice timbre, body pose, and scene layout in joint audio-video generation.