Article
Effect of frequency domain sparsity on musical scene analysis abilities of hearing-impaired listeners (en)
* Presenting author
Abstract:
Music pre-processing methods have become a popular area of research with the goal to make music more accessible to listeners with sensorineural hearing impairment (HI). Our previous study showed that HI preferred spectrally sparser multi-track mixes. Nevertheless, the acoustical basis of transparency of mixes for HI remains poorly understood. Here, we attempt to assess the ability to detect a musical target within mixes with varying spectral sparsity. In our experiment, 30 normal-hearing (NH) and 24 HI participants with predominantly moderate to severe hearing loss were tested. The target that was to be detected in the mixes was from the categories lead vocals, bass guitar, drums, guitar, and piano. Our results show that overall performance depreciated significantly with increasing hearing loss. This effect was especially prominent for lead vocals and bass guitar targets. Notably, there was an improvement of accuracy among those participants with higher levels of hearing impairment for lead vocals, drums, and piano targets when the mixes were spectrally sparser. These findings suggest that hearing loss plays a pivotal role for musical scene analysis abilities and frequency domain sparsity in music may mitigate some of the discrepancies brought on by hearing loss.