Abstract:

Reduced frequency selectivity as occurs in the cochlear impaired ear has been shown to be a main factor for degraded speech perception [J. M. Festen and R. Plomp, J. Acoust. Soc. Am. 76, 652--662 (1983)]. It is also assumed to have negative effects on music perception. In this paper the hypothesis is tested that the loudness relations between spectral peaks and nonmasked lower-level components, which depend on the individual frequency selectivity, are important for (a) subjective sound quality (listening comfort) and (b) clarity, as measured in terms of detectability of altered notes in music excerpts. In order to test this hypothesis a signal processing algorithm has been utilized which enhances or suppresses the lower-level spectral components of a complex signal according to a masking function. Processed signals have been presented to normal-hearing and cochlear-impaired subjects. Results revealed that subjects with reduced frequency selectivity---as measured with psychoacoustical tuning curves---tend to prefer music signals with enhanced lower-level components. In the clarity experiment, results obtained with hearing-impaired subjects show improved detectability with suppression of the lower-level components (increase of spectral contrasts). [Work supported by Austrian Academy of Sciences.]