ReverbFX: A Dataset of Room Impulse Responses Derived from Reverb Effect Plugins for Singing Voice Dereverberation

TL;DR

ReverbFX is a novel dataset of artificial room impulse responses derived from reverb plugins, enabling improved singing voice dereverberation research and model training.

Contribution

The paper introduces ReverbFX, a new dataset of plugin-based RIRs, and demonstrates its effectiveness in training models for singing voice dereverberation.

Findings

Models trained on ReverbFX outperform those trained on real RIRs in artificial reverberation scenarios.

ReverbFX enables better benchmarking of dereverberation methods for singing voices.

The dataset facilitates research using common music production reverb effects.

Abstract

We present ReverbFX, a new room impulse response (RIR) dataset designed for singing voice dereverberation research. Unlike existing datasets based on real recorded RIRs, ReverbFX features a diverse collection of RIRs captured from various reverb audio effect plugins commonly used in music production. We conduct comprehensive experiments using the proposed dataset to benchmark the challenge of dereverberation of singing voice recordings affected by artificial reverbs. We train two state-of-the-art generative models using ReverbFX and demonstrate that models trained with plugin-derived RIRs outperform those trained on realistic RIRs in artificial reverb scenarios.