A modeling and algorithmic framework for (non)social (co)sparse audio restoration
This work provides an incremental improvement in audio restoration techniques for practitioners working with denoising and declipping.
This paper proposes a unified framework for audio restoration using various sparse priors, including analysis and synthesis, and regular and structured sparsity. It demonstrates its effectiveness in denoising and declipping scenarios, achieving speedups of 20% or more with analysis sparse priors and substantial quality in declipping.
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.