Stochastic Restoration of Heavily Compressed Musical Audio using Generative Adversarial Networks

TL;DR

This paper explores using a stochastic GAN-based approach to restore heavily compressed musical audio, demonstrating improved quality over standard MP3 compression at low bitrates and showing the stochastic model's advantage over deterministic methods.

Contribution

It introduces a stochastic GAN architecture for musical audio restoration, providing a novel approach to improve quality of heavily compressed audio signals.

Findings

Models improve audio quality over MP3 at 16 and 32 kbit/s.

Stochastic generators produce outputs closer to original signals than deterministic ones.

Extensive evaluation with objective metrics and listening tests supports the effectiveness.

Abstract

Lossy audio codecs compress (and decompress) digital audio streams by removing information that tends to be inaudible in human perception. Under high compression rates, such codecs may introduce a variety of impairments in the audio signal. Many works have tackled the problem of audio enhancement and compression artifact removal using deep learning techniques. However, only a few works tackle the restoration of heavily compressed audio signals in the musical domain. In such a scenario, there is no unique solution for the restoration of the original signal. Therefore, in this study, we test a stochastic generator of a Generative Adversarial Network (GAN) architecture for this task. Such a stochastic generator, conditioned on highly compressed musical audio signals, could one day generate outputs indistinguishable from high-quality releases. Therefore, the present study may yield insights into more efficient musical data storage and transmission. We train stochastic and deterministic generators on MP3-compressed audio signals with 16, 32, and 64 kbit/s. We perform an extensive evaluation of the different experiments utilizing objective metrics and listening tests. We find that the models can improve the quality of the audio signals over the MP3 versions for 16 and 32 kbit/s and that the stochastic generators are capable of generating outputs that are closer to the original signals than those of the deterministic generators.