MMAudioReverbs: Video-Guided Acoustic Modeling for Dereverberation and Room Impulse Response Estimation

TL;DR

This paper introduces MMAudioReverbs, a framework that leverages pretrained video-to-audio models to perform dereverberation and RIR estimation, bridging visual cues with physical room acoustics.

Contribution

It proposes a novel method to utilize existing V2A models for physically grounded room-acoustic processing without changing network architecture.

Findings

Audio and visual cues have different advantages depending on room acoustics.

V2A models can be fine-tuned for dereverberation and RIR estimation.

The approach works with small datasets.

Abstract

Although recent video-to-audio (V2A) models excelled at synthesizing semantically plausible sounds from visual inputs, they do not explicitly model room-acoustic effects such as reverberation or room impulse responses (RIRs), and thus offer limited controllability over these effects. However, we hypothesize that such V2A models implicitly have semantic knowledge of the relationship between spatial audio and the corresponding vision cues. In this paper, we revisit a V2A model for the sake of the above, and propose the way to utilize the pretrained model as prior for physically grounded room-acoustic processing. Based on one of the state-of-the-art V2A models, MMAudio, we propose MMAudioReverbs that is a unified framework dealing with i) dereverberation and ii) room impulse response (RIR) estimation without network architectural modification, and fine-tuned on a small dataset. Experimental results showed that audio and visual cues respectively have advantage depending on the type of physical room acoustics. It implies that foundation V2A models can be used for physically grounded room-acoustic analysis.