Frequency selectivity and consonant recognition for hearing-impaired and normal-hearing listeners with equivalent masked thresholds.

Thresholds in notched-noise maskers (NN) and narrow-band maskers (NB) were measured for normal-hearing and hearing-impaired subject pairs who were listening in a background of spectrally shaped broadband noise (SSBB). Consonant recognition was also measured in SSBB. SSBB was adjusted so that thresholds in that noise for each normal-hearing/hearing-impaired subject pair were equal. Threshold and signal-level differences between subject pairs were minimized with the addition of threshold-elevating SSBB, and the presence of a noise background for both groups provided a comparable listening environment for all subjects. At signal frequencies outside masker passbands, thresholds in NN and NB for hearing-impaired subjects were higher than for normal-hearing subjects, although threshold differences were much smaller than observed between normal-hearing and hearing-impaired subjects without SSBB. No consistent differences in consonant recognition measured in SSBB were observed between groups. The pattern of results is comparable to that observed in a previous experiment [Dubno and Schaefer, J. Acoust. Soc. Am. 91, 2110-2121 (1992)] in which thresholds in NN and NB, and consonant recognition, for hearing-impaired listeners were compared to results obtained for normal-hearing subjects, but only normal-hearing subjects listened in SSBB. Using a modified power-law model of masking additivity, thresholds in combined-masker conditions were estimated. Masking effects for spectrally overlapping maskers were similar for normal-hearing and hearing-impaired listeners and suggest that residual differences between subject groups are not due to the presence of an additional background noise.