Cellular Network Speech Enhancement: Removing Background and Transmission Noise

TL;DR

This paper presents a deep learning-based speech enhancement method tailored for VoIP applications like Google Meet, achieving state-of-the-art results in reducing background and transmission noise to improve speech clarity.

Contribution

The work introduces a novel approach that specifically addresses transmission noise in VoIP calls, outperforming existing industrial standards in speech enhancement metrics.

Findings

Achieved 1.92 PESQ score indicating high speech quality

Attained 0.88 STOI demonstrating improved intelligibility

Surpassed industrial performance benchmarks in VoIP noise suppression

Abstract

The primary objective of speech enhancement is to reduce background noise while preserving the target's speech. A common dilemma occurs when a speaker is confined to a noisy environment and receives a call with high background and transmission noise. To address this problem, the Deep Noise Suppression (DNS) Challenge focuses on removing the background noise with the next-generation deep learning models to enhance the target's speech; however, researchers fail to consider Voice Over IP (VoIP) applications their transmission noise. Focusing on Google Meet and its cellular application, our work achieves state-of-the-art performance on the Google Meet To Phone Track of the VoIP DNS Challenge. This paper demonstrates how to beat industrial performance and achieve 1.92 PESQ and 0.88 STOI, as well as superior acoustic fidelity, perceptual quality, and intelligibility in various metrics.