Window Size Versus Accuracy Experiments in Voice Activity Detectors

TL;DR

This study evaluates how window size affects the accuracy of three voice activity detection algorithms across real-world audio streams, providing insights for optimizing VAD system performance.

Contribution

It offers a comparative analysis of VAD algorithms with different window sizes and explores hysteresis effects, which is a novel practical evaluation.

Findings

Silero outperforms WebRTC and RMS in accuracy.

Hysteresis improves WebRTC's performance.

Optimal window size varies across algorithms.

Abstract

Voice activity detection (VAD) plays a vital role in enabling applications such as speech recognition. We analyze the impact of window size on the accuracy of three VAD algorithms: Silero, WebRTC, and Root Mean Square (RMS) across a set of diverse real-world digital audio streams. We additionally explore the use of hysteresis on top of each VAD output. Our results offer practical references for optimizing VAD systems. Silero significantly outperforms WebRTC and RMS, and hysteresis provides a benefit for WebRTC.