Effects of Loudspeaker Configuration on Speech-in-Noise Evaluation of Remote Microphone Systems

Purpose: This study aimed to examine how loudspeaker configuration influences the evaluation of personally worn remote microphone (RM) systems under a single signal-to-noise ratio (SNR) condition. Method: This study included 10 school-aged children with hearing loss who used hearing devices and RM systems. Speech recognition was assessed using a Japanese monosyllabic word test at a nominal SNR of −5 dB. Three loudspeaker configurations were tested: unilateral noise at 45°, unilateral noise at 0°, and bilateral uncorrelated noise at ±45°. Each condition was evaluated under microphone only (MIC) and RM input modes. Speech recognition accuracy was analyzed using a generalized linear mixed-effects model, with input condition and loudspeaker configuration as fixed effects and participant and word items as random effects. Results: The use of RM significantly increased the probability of correct speech recognition across all loudspeaker configurations. RM performance was consistent across configurations, with no significant interaction between input condition and loudspeaker configuration. Conclusions: RM systems provided a relatively consistent benefit in speech recognition across different loudspeaker configurations. Meanwhile, speech recognition with MIC input was highly sensitive to noise arrangement. Therefore, the effectiveness of RM systems is best characterized by their ability to stabilize performance under adverse listening conditions rather than by condition-specific enhancement. A unilateral 45° noise configuration appears to provide a practical and feasible setup for clinical RM evaluation. Supplemental Material: https://doi.org/10.23641/asha.34113570