The effects of "noise suppression" hearing aids on consonant recognition in speech-babble and low-frequency noise.

We evaluated the performance of experienced hearing-aid users wearing seven different commercially available "noise-suppression" hearing aids. Two hearing aids, the Audiotone A-54 and the Telex 363C, used amplitude compression. The others, two versions of a Maico hearing aid SP147, a Richards ASE-B, a Rion HB-69AS, and a Siemens 283ASP, are designed to attenuate specific frequency regions in the presence of noise. Sixteen subjects listened to 13 consonants in the form (i)-consonant-(i) with six replications per consonant (78 items). Performance was measured with the compression or noise-suppression circuit on and off in the presence of speech-babble noise and in continuous low-frequency noise. Measurements were also obtained with the suppression circuit "off" but without any background noise. The results suggested that only a few subjects benefitted from the noise-suppression circuits, and in several cases performance in noise was poorer with the noise suppression circuit than without it. An information-transfer analysis of the errors indicated that enhanced or decreased performance was generally a result of changes across all phonetic features, not specific ones.