Speech discrimination with an 8-channel compression hearing aid and conventional aids in background of speech-band noise

Abstract

The design for a multichannel compression hearing aid was developed from previous experimental and theoretical work in our laboratory concerning pitch perception in normal-hearing subjects. The new hearing aid, implemented with off-line digital signal processing, was tested on twenty subjects with sensorineural hearing loss using speech sounds in a background of speech-spectrum noise. Five signal-to-noise ratios (+15 to -5dB) were used at two noise levels (60 and 70 dB SPL). Hearing-loss subjects listened to these stimuli under three different conditions: a) processed by the new multichannel compression hearing aid; b) processed by a conventional hearing aid; and c) unprocessed. The performance of normal-hearing subjects with the unprocessed stimuli provided another condition against which the performance in the two hearing aid conditions could be evaluated. Both aided conditions provided improved performance over the unprocessed condition and the multichannel compression aid produced better performance than the conventional aid. In the case of 4 of the 20 subjects, with less severe gradually sloping hearing losses, the new multichannel compression aid produced near-normal performance even at low signal-to-noise ratios. Some aspects of the results also suggested that learning to use the aid was more important in the case of the multichannel compression aid than in the case of the conventional aid. These results indicate that a multichannel compression hearing aid can be very effective in some individuals with sensorineural hearing loss and is superior to a conventional hearing aid in most subjects.