Study of multichannel amplitude compression and linear amplification for persons with sensorineural hearing loss

Abstract

Experiments were conducted on five listeners with sensorineural hearing impairments using two 16-channel, computer-controlled, amplitude compression systems, and four linear systems. One of the compression systems was designed to restore normal equal loudness contours, the other employed reduced high-frequency emphasis and reduced compression ratios. The linear systems differed only in their frequency-gain characteristics (orthotelephonic plus three characteristics with high-frequency emphasis that were expected to produce better results than orthotelephonic). In the main experiment, all systems were compared for each listener using nonsense CVC monosyllables and sentence materials spoken by male and female talkers and presented in quiet/anechoic and noisy/reverberant environments at the most comfortable level for each listener. The linear systems with high-frequency emphasis performed substantially better than the orthotelephonic system. Performance with compression was generally slightly worse than with linear amplification. Compression was superior to linear amplification only when speech materials with significant item-to-item level variation were used in quiet with subjects with more severe losses and when reduced input speech levels were used. To the extent that these two conditions represent real-life communication conditions, these results suggest that compression is preferable to linear amplification in a wearable hearing aid.