Speech recognition thresholds in temporally complex backgrounds: Effects of hearing loss and noise masking

When listening to speech in a fluctuating background, subjects with normal hearing can operate at a higher signal-to-noise ratio (SNR) by taking advantage of momentary decreases in background level. Subjects with even a mild hearing loss are less able to do so. This was examined here in subjects with normal hearing, with sensorineural hearing loss, and with normal hearing whose pure-tone thresholds were elevated via masking to match those of the hearing-impaired subjects. Four backgrounds were used: speech-shaped nose (SS); speech-shaped noise modulated by the envelope of a single talker (ST); speech-shaped noise modulated by the envelope of a multi-talker babble (MT); and speech-shaped noise modulated by a 10-Hz square wave at a depth of 100% (SQ). The rms level of the background was 70 dB SPL. The level of the sentences was varied adaptively to obtain a recognition threshold SNR. For the normal-hearing subjects, the average SNRs in the ST and SQ backgrounds were 3 to 8 dB lower, respectively, than those for either the SS or MT backgrounds. There generally was less ‘‘masking release’’ for the hearing-impaired and noise-masked subjects. For some hearing-impaired subjects, the release was smaller than that in their noise-masked counterpart, suggesting the audibility alone may not always account for the reduced masking release observed in the hearing impaired. [Work supported by NIDCD.]