Evaluation of elevated hearing thresholds produced by models for the simulation of cochlear hearing loss

Introduction: As the statistical distribution of hearing thresholds moves towards stronger hearing impairments with age, plausible simulations of its perceptual effects are useful for a variety of applications. Four software tools for the simulation of hearing loss, capable of generating audible output, were evaluated. An overview of the simulators is presented, their capabilities and significant signal processing steps are introduced. Methods: The perceptual evaluation focuses on the simulation of elevated hearing thresholds. Two listening experiments were conducted to assess how accurately the hearing loss simulators can reproduce target audiograms, i.e., elevated hearing thresholds when normal-hearing listeners are subjected to the simulation. Mild and moderate degrees of simulated hearing loss conditions were defined based on typical hearing thresholds of 70- and 80-year-olds. The complementary technical analysis addresses additional simulated consequences of cochlear hearing loss by investigating input-vs.-output level functions and spectral smearing effects. Results: Statistically significant differences between simulators were found: Good agreement with the target hearing thresholds was found for the simulator “WHIS” (deviations 0 to 6 dB), while the others showed deviations of varying degree (−29 to 7 dB). The created input-output functions proved to be suitable for demonstrating expansive dynamic range processing and explaining the listening experiment results.