Sparsity-based audio declipping methods: overview, new algorithms, and large-scale evaluation

05/19/2020
by   Clément Gaultier, et al.
0

Recent advances in audio declipping have substantially improved the state of the art in certain saturation regimes. Yet, practitioners need guidelines to choose a method, and while existing benchmarks have been instrumental in advancing the field, larger-scale experiments are needed to guide such choices. First, we show that the saturation levels in existing small-scale benchmarks are moderate and call for benchmarks with more perceptually significant saturation levels. We then propose a general algorithmic framework for declipping that covers existing and new combinations of flavors of state-of-the-art techniques exploiting time-frequency sparsity: synthesis vs analysis sparsity, with plain or structured sparsity. Finally, we systematically compare these combinations and state-of-the-art methods. Using a large-scale numerical benchmarks and a smaller scale formal listening test, we provide guidelines for various saturation levels, both for speech and various musical "genres" from the RWC database. The code is made publicly available for the purpose of reproducible research and benchmarking.

READ FULL TEXT
research
10/31/2018

Introducing SPAIN (SParse Audion INpainter)

A novel sparsity-based algorithm for audio inpainting is proposed by tra...
research
01/24/2023

Mesostructures: Beyond Spectrogram Loss in Differentiable Time-Frequency Analysis

Computer musicians refer to mesostructures as the intermediate levels of...
research
05/05/2021

How do Voices from Past Speech Synthesis Challenges Compare Today?

Shared challenges provide a venue for comparing systems trained on commo...
research
08/10/2017

Distributed Training Large-Scale Deep Architectures

Scale of data and scale of computation infrastructures together enable t...
research
06/09/2021

Time-Frequency Phase Retrieval for Audio – The Effect of Transform Parameters

In audio processing applications, phase retrieval (PR) is often performe...
research
01/08/2022

Evaluation of histological findings with severity grade, to analyze toxicology in-vivo studies

In-vivo toxicological studies are characterized by multiple primary endp...
research
03/01/2021

TSSOS: a Julia library to exploit sparsity for large-scale polynomial optimization

The Julia library TSSOS aims at helping polynomial optimizers to solve l...

Please sign up or login with your details

Forgot password? Click here to reset