Initial stop voicing perception for cue‐adjusted syllables for impaired‐ and normal‐hearing listeners

Abstract

Further work will be presented on voicing cues, specifically for initial stops, used by the hearing‐impaired. The set of test syllables /pæd, kæd, tæd, bæd, gæd, dæd/ was randomized 10 times on a list recorded by a male. Based on acoustical analyses, the potential sources of voicing cues in the syllables were: release‐burst duration, transition‐pattern of vowel onset, and vowel pitch. Among different conditions these cues in the syllables were modified cumulatively, so as to eliminate the contribution to voicing of each cue. The modifications were made through LPC resynthesis and/or waveform deletions. The conditions were administered monaurally at comfortable levels to six normal‐ and 21 impaired‐hearing adults. The syllables of each condition were presented in single‐interval identification trials without feedback. Generally, a similar pattern of results was seen between the impaired and the normal groups, although the impaired group performed somewhat poorer. Perception of initial stop voicing was unaffected by equalized vowel pitch. For voiceless stops deletion of the release burst reduced voicing perception for most listeners. However, for voiced stops the burst deletion worsened performance for less than half of the listeners. When the voiced burst and vowel‐onset‐transitions were removed, voicing perception was reduced for most listeners.