Speech-in-noise recognition during hearing protector use: Human performance and acoustic prediction

Hearing protection devices (HPDs) are essential safety equipment for workers in noisy settings but can also degrade auditory perception. Some reports have suggested that HPDs can affect speech intelligibility, particularly at low levels, while others reported no effect for normal-hearing listeners. Here, we investigated the effects of different HPDs on speech intelligibility at various signal-to-noise ratios (SNRs) using the Quick Speech-In-Noise (QuickSIN) test and the Modified Rhyme Test (MRT). 124 audiometrically normal subjects were tested across 2 study sites in Open Ear and 6 HPD listening conditions. Target levels were fixed at 70 dBA, while noise levels varied (QuickSIN: multi-talker babble co-located with the target; MRT: spatially distributed pink noise). Subject performance was comparable to Open Ear for most HPD conditions; however, performance was degraded in both tasks by a high insertion-loss passive earplug, and modestly improved in MRT by a pair of electronic earmuffs. Behavioral performance was compared to predictions using the Speech Intelligibility Index (SII) calculated in two ways: first, using average speech and noise stimulus levels, second, by filtering those stimulus levels with appropriate head-related transfer functions (HRTFs) of a GRAS 45CB acoustic manikin to incorporate the spatial dependence of each sound source in the ear canal (sSII). Comparisons to subject performance reveal a substantial increase in predictive performance with sSII (R²=0.497) over SII (R²=0.212) for MRT. Results suggest speech perception with most HPDs is comparable to Open Ear, and that incorporating spatial information improves acoustical predictions.