Segmental interference by the masker modulation spectrum in sentences: Effects of age and hearing loss

Abstract

Younger normal-hearing, older normal-hearing, and older hearing-impaired subjects listened to sentences with consonants or vowels replaced with noise modulated by the envelope of a competing talker. The modulation spectrum of the noise was low-pass filtered at different cutoff frequencies. Sentences were spectrally shaped according to listeners’ individual audiometric thresholds to ensure sufficient audibility; in addition, a second group of younger normal-hearing subjects listened to sentences that were spectrally shaped according to the mean audiogram of the hearing-impaired subjects. Results demonstrated declines in sentence intelligibility for older as compared to younger listeners, with more evident declines for older hearing-impaired listeners. Declines in sentence intelligibility based on vowel cues with increasing modulation rates in the competing noise were noted for all listener groups. In contrast, sentence intelligibility based on consonant cues increased when the competing noise included modulation rates between 8 and 16 Hz. An adverse effect of spectral shaping was also observed, with an overall decrease in performance for younger spectrally matched listeners. Thus, in temporally complex listening conditions, spectral shaping to assure audibility may not fully control for reduced sentence intelligibility among hearing-impaired listeners and may interact with temporal properties of the speech signal. [Work supported by NIH/NIDCD and ASHA.]