Listening to Sounds of Silence for Speech Denoising

TL;DR

This paper presents a deep learning approach for speech denoising that exploits incidental silent intervals in speech signals to effectively learn noise characteristics and improve denoising performance across languages.

Contribution

The study introduces a novel speech denoising method that leverages silent intervals to learn noise dynamics, outperforming existing methods and demonstrating strong generalization.

Findings

Outperforms several state-of-the-art denoising methods

Effectively generalizes to unseen spoken languages

Utilizes silent intervals to learn noise features

Abstract

We introduce a deep learning model for speech denoising, a long-standing challenge in audio analysis arising in numerous applications. Our approach is based on a key observation about human speech: there is often a short pause between each sentence or word. In a recorded speech signal, those pauses introduce a series of time periods during which only noise is present. We leverage these incidental silent intervals to learn a model for automatic speech denoising given only mono-channel audio. Detected silent intervals over time expose not just pure noise but its time-varying features, allowing the model to learn noise dynamics and suppress it from the speech signal. Experiments on multiple datasets confirm the pivotal role of silent interval detection for speech denoising, and our method outperforms several state-of-the-art denoising methods, including those that accept only audio input (like ours) and those that denoise based on audiovisual input (and hence require more information). We also show that our method enjoys excellent generalization properties, such as denoising spoken languages not seen during training.