Effects of some acoustic-cue modifications on the perception of voiced and unvoiced final stop consonants for hearing-impaired listeners

Abstract

Various acoustic cues to voicing were modified systematically in a corpus of ten utterances each of dap, dak, dat, dab, dag, and dad spoken by a male. The acoustic cues, modified separately and/or simultaneously, were: vowel duration, closure duration, murmur presence, final stop-burst presence, and vowel offset. The modifications were accomplished by either deleting, iterating, or transposing segments of digital representations of the utterances, using waveform manipulations via computer. The digital stimuli, with and without acoustic-cue modifications, were presented to 23 hearing-impaired listeners for identification of the final consonant within the closed set of six stimuli. In general, different acoustic-cue modifications degraded the perception of voicing among the listeners. Burst and/or murmur absence greatly affected performance for some listeners, but not for others. Alterations to vowel duration degraded voicing perception for a few listeners. The elimination of the last ten vowel pitch periods, with burst and murmur also absent, degraded voicing perception for most listeners. Adjustments to closure duration had little effect.