Improving Deep Speech Denoising by Noisy2Noisy Signal Mapping

TL;DR

This paper introduces a self-supervised deep learning method for speech denoising that does not require clean training signals, using noisy realizations of the same speech for training, and demonstrates its effectiveness over traditional supervised methods.

Contribution

A novel self-supervised deep speech denoising approach that eliminates the need for clean training data by leveraging noisy signal pairs.

Findings

Outperforms conventional supervised denoising methods on multiple metrics

Effective in real-world noisy environments

Validated through extensive experiments and field tests

Abstract

Existing deep learning-based speech denoising approaches require clean speech signals to be available for training. This paper presents a deep learning-based approach to improve speech denoising in real-world audio environments by not requiring the availability of clean speech signals in a self-supervised manner. A fully convolutional neural network is trained by using two noisy realizations of the same speech signal, one used as the input and the other as the output of the network. Extensive experimentations are conducted to show the superiority of the developed deep speech denoising approach over the conventional supervised deep speech denoising approach based on four commonly used performance metrics and also based on actual field-testing outcomes.