Authors: Luuk Schipper¹; Nouri Khalass²; Deniz Başkent³; Eleanor E. Harding⁴; Etienne Gaudrain⁵
¹Rijksuniversiteit Groningen; ²Golden Hearing B.V.; ³University Medical Center Groningen; ⁴University of Groningen; ⁵Lyon Neuroscience Research Center
Background: Cochlear implant (CI) users experience degraded music quality. Golden Hearing has developed an algorithm, the AM-Booster, for hearing aid users aiming to enhance music appreciation by amplifying the magnitude of amplitude modulation (AM) in the signal. While effective for hearing aid users, the optimal parameters for CI users are not yet known. The parameter combination providing the best experience might be determined by specific musical characteristics, such as genre or pre-existing modulations. This study investigates optimal AM-Booster parameters for CI users and their relationship to musical genre.
Method: 150 musical excerpts (10 s) from 15 genres were selected from the Free Music Archive. After level adjustment and filtering for quality, the corpus will be processed through a CI-processor with the AM-Booster algorithm. The resulting electrical stimulation will be examined to evaluate the Booster’s effect. Based on previous work, two parameters known to influence AM most significantly will be sampled with fine granularity, while others will undergo sparse adjustments.
Results: The excerpts will be clustered based on their modulation spectra to inspect the relationship to genre. This analysis will be used to determine usable parameter settings. In a planned follow-up study, these settings will undergo quality ratings by CI users to verify if the AM-Booster improves the music listening experience.
Conclusion: Determining optimal AM-Booster parameters based on the modulation characteristics of different musical genres may lead to personalized signal processing strategies. This approach aims to bridge the gap in music appreciation for cochlear implant users by tailoring algorithm settings to the specific acoustic properties of the music being heard.
