ADNAC: Audio Denoiser using Neural Audio Codec

TL;DR

This paper introduces ADNAC, a neural audio codec-based approach for music denoising that leverages a large, diverse dataset and a multi-objective loss function to improve audio restoration quality.

Contribution

It adapts a neural audio codec for denoising, overcoming traditional architecture limitations with a large dataset and multi-faceted training objectives.

Findings

Effective denoising demonstrated on diverse audio samples

Improved fidelity over traditional methods

Proof-of-concept for high-quality audio restoration

Abstract

Audio denoising is critical in signal processing, enhancing intelligibility and fidelity for applications like restoring musical recordings. This paper presents a proof-of-concept for adapting a state-of-the-art neural audio codec, the Descript Audio Codec (DAC), for music denoising. This work overcomes the limitations of traditional architectures like U-Nets by training the model on a large-scale, custom-synthesized dataset built from diverse sources. Training is guided by a multi objective loss function that combines time-domain, spectral, and signal-level fidelity metrics. Ultimately, this paper aims to present a PoC for high-fidelity, generative audio restoration.