New results on multichannel amplitude compression for the hearing impaired

Abstract

Experiments were conducted on five listeners with sensorineural impairments and reduced dynamic ranges using 16‐channel computer controlled, amplitude compression systems. Each subject was tested with two compression systems and, for reference purposes, four linear systems. One of the compression systems was chosen to restore normal equal loudness contours; the other employed reduced high‐frequency emphasis and reduced compression ratios. The four linear systems differed only in the frequency‐gain characteristic (orthotelephonic plus three characteristics with high‐frequency emphasis that were expected to produce better results than orthotelephonic). The six systems were compared on each of the five subjects using nonsense CVC monosyllables and sentence material spoken by male and female talkers and presented in quiet/anechoic and noisy/reverberant environments at the most comfortable level for each listener. The principal results of these tests were: (1) The linear system with high‐frequency emphasis produced substantially better performance than the orthotelephonic system; and (2) neither of the two compression systems led to significantly better performance than the best linear system. In considering these results, it should be noted that (a) the long‐term level of the speech material was held constant before processing, (b) the subjects suffered from only moderate losses, and (c) only two compression systems were tested. If one takes account of the differences in the linear systems to which the compression systems were compared, these results are consistent with those of Villchur [J. Acoust. Soc. Am. 53, 1646–1657 (1973)], and of Barfod [Reports No. 4 and 11, Acoust. Lab., Tech. Univ., Denmark (1972, 1976)]. [Work supported by NIH Grants 5 R01 NS12846 and 5 T01 GM01555.]