01whole.pdf (3.72 MB)
Hearing aids and music
thesisposted on 2022-03-28, 10:05 authored by Rémi Marchand
Currently, hearing aids (HAs) are adjusted to compensate for individual hearing loss primarily to maximise the clarity and comfort of speech. Electroacoustic characteristics and settings of HAs may be ideal for speech recognition, but not for music enjoyment. The aims of this study were to better understand the musical listening habits of HA users, to identify the main issues they experience while listening to music and to develop signal processing recommendations specifically for music. A survey on music listening with HAs was conducted and 151 respondents were recruited. The survey showed that HA users listen mainly to recorded music at home and use the HAs in their default program. 30% of the respondents were dissatisfied with the performance of their HAs and experience problems that need to be addressed. The most prevalent problems identified were related to the HA sound quality and suggest that frequency-specific gains and compression algorithms should be improved. A follow-up study involved a controlled listening experiment to further understand the signal processing strategies preferred by participants when listening to instrumental music in relation to the problems highlighted by the survey. Manipulations of the signal processing comprise changes in frequency-specific insertion gains and compression ratios relative to those prescribed by the NAL-NL2 standard fitting formula. The preferred frequency-specific amplification provided by the HAs indicates that more gain at low frequency and less gain at high frequency than prescribed byNAL-NL2 standard fitting formulae is needed. The preferred amount of compression applied by the HAs depended on the music stimuli at the input of the HA with overall preference for more compression than prescribed by NAL-NL2.Future research should derive amplification prescription schemes specifically targeted at optimizing the enjoyment of music with HAs. The optimal prescription will potentially depend on the acoustic characteristics of the specific music being listened to.