ThinkSound: Chain-of-Thought Reasoning in Multimodal Large Language Models for Audio Generation and Editing

TL;DR

ThinkSound introduces a chain-of-thought reasoning framework for interactive, stepwise audio generation and editing in videos, combining multimodal reasoning with a new dataset to improve fidelity and controllability.

Contribution

It presents a novel CoT-based multimodal framework and dataset for structured reasoning in audio generation and editing from videos.

Findings

Achieves state-of-the-art video-to-audio generation performance.

Demonstrates effective interactive object-centric refinement.

Excels in out-of-distribution audio generation benchmarks.

Abstract

While end-to-end video-to-audio generation has greatly improved, producing high-fidelity audio that authentically captures the nuances of visual content remains challenging. Like professionals in the creative industries, this generation requires sophisticated reasoning about items such as visual dynamics, acoustic environments, and temporal relationships. We present ThinkSound, a novel framework that leverages Chain-of-Thought (CoT) reasoning to enable stepwise, interactive audio generation and editing for videos. Our approach decomposes the process into three complementary stages: foundational foley generation that creates semantically coherent soundscapes, interactive object-centric refinement through precise user interactions, and targeted editing guided by natural language instructions. At each stage, a multimodal large language model generates contextually aligned CoT reasoning that guides a unified audio foundation model. Furthermore, we introduce AudioCoT, a comprehensive dataset with structured reasoning annotations that establishes connections between visual content, textual descriptions, and sound synthesis. Experiments demonstrate that ThinkSound achieves state-of-the-art performance in video-to-audio generation across both audio metrics and CoT metrics, and excels in the out-of-distribution Movie Gen Audio benchmark. The project page is available at https://ThinkSound-Project.github.io.