Consonant recognition performance of hearing-impaired listeners using one linear and three nonlinear hearing aids

Consonant recognition performance was examined in 18 subjects listening in quiet and speech noise with linear and nonlinear hearing aids. Subjects were divided into three groups by audiometric configuration: flat, moderately sloping, and sharply sloping. Nonlinear amplification schemes included adaptive high-pass filtering (Argosy's Manhattan II, experimentally modified; MANe), MANe circuitry followed by expansion (EXP), and infinite amplitude clipping (IAC). Consonant-to-vowel ratios (CVRs) were calculated for syllables processed through each hearing aid. Performance with the IAC was significantly poorer than with the other amplifiers across audiograms in noise. Intersubject variability in performance was high, even within audiogram groups. High-frequency phonemes were more often audible with the EXP than with the other hearing aids for subjects with moderately and severely sloping audiograms. Output CVRs increased for some phonemes with a nonlinear hearing aid versus linear, but recognition of individual phonemes did not correlate significantly with CVR.