On the Use of Wide Dynamic Range Compression and Other Algorithms to Improve Hearing Protection of Workers with Hearing Impairment: A Preliminary Study on Speech Intelligibility

Abstract

Despite being affected by one of the most common occupational diseases worldwide, workers with hearing loss remain without any consensual solutions when it comes to protecting their residual hearing while working safely and efficiently. One solution to be explored is to adapt signal processing algorithms commonly used in hearing aids for hearing protection applications and to implement them on a research platform for evaluation. The algorithms to be assessed include Wide Dynamic Range Compression (WDRC) to amplify sound based on the user’s hearing loss while protecting their hearing, Modulation-Based Digital Noise Reduction (MBDNR) to enhance speech, and an Automatic Gain Control (AGC) to protect from excessive exposure. A major need for workers is the ability to communicate in noise. Different parameters and algorithms combination have been proven to affect speech intelligibility in hearing aid devices. To select the proper combination and parameters before hardware implementation, this study explores the impact of different algorithm configurations on speech intelligibility in factory noise in a laboratory environment. Participants either without hearing impairment or with mild to moderate hearing loss undergo five different Hearing in Noise Test (HINT) in order to measure their Speech Reception Thresholds (SRTs) using the MATLAB Speech Test Environment (MSTE). Differences in the SRTs obtained when simulating different hearing devices and other metrics are analyzed to highlight a potential benefit of the signal processing algorithms in terms of speech intelligibility improvement in a typical factory noise.