Accuracy of audiometric test room simulations of three real‐world listening environments

Abstract

Hearing aid benefit depends primarily on the extent to which amplification facilitates speech understanding in typical everyday listening environments. In the hearing aid fitting process, improved speech understanding is often measured in an audiometric test room. However, because audiometric test rooms are smaller, quieter, and less reverberant than typical rooms, these data may not accurately predict speech understanding in daily life. This study was undertaken to evaluate the validity of three simulated real‐world listening environments created in an audiometric test room. The three environments represented a typical living room, cocktail party, and classroom, respectively. Twenty normal‐hearing subjects, listening monaurally, provided intelligibility scores for four phonetic contrasts produced by each of three normal talkers. Intelligibility obtained in the real environment was compared with that measured in the corresponding simulated environment. Results indicated that the relative intelligibility of talkers and phonetic contrasts remained essentially constant across each real‐simulated environment pair, and that significant talker × contrast interactions seen in the real environments were usually reproduced in the simulated environments. However, there were somewhat fewer significant intelligibility differences in the simulated environments than in the real environments. Also, the intelligibility of one talker deteriorated more than expected in the simulated reverberant environment. Overall, the outcome suggested that these typical listening environments were rather accurately simulated (for monaural listening) in an audiometric test room using appropriate adjustments of presentation level, signal‐to‐babble ratio, and synthetic reverberation effects.