A modeling and algorithmic framework for (non)social (co)sparse audio restoration

TL;DR

This paper introduces a comprehensive framework for audio restoration that integrates analysis and synthesis sparse priors, demonstrating improved speed and quality in denoising and declipping tasks through extensive experiments.

Contribution

It presents a unified modeling and algorithmic framework that combines various sparsity priors and structured shrinkage operators for audio restoration.

Findings

Analysis sparse prior speeds up processing by over 20%.

Both social and plain sparsity priors achieve high declipping quality.

Performance varies between priors depending on the task (denoising or declipping).

Abstract

We propose a unified modeling and algorithmic framework for audio restoration problem. It encompasses analysis sparse priors as well as more classical synthesis sparse priors, and regular sparsity as well as various forms of structured sparsity embodied by shrinkage operators (such as social shrinkage). The versatility of the framework is illustrated on two restoration scenarios: denoising, and declipping. Extensive experimental results on these scenarios highlight both the speedups of 20% or even more offered by the analysis sparse prior, and the substantial declipping quality that is achievable with both the social and the plain flavor. While both flavors overall exhibit similar performance, their detailed comparison displays distinct trends depending whether declipping or denoising is considered.