Comparison of different voice samples for perceptual analysis.

Choice of voice sample material can influence perceptual judgments by a jury. The twofold purpose of this study was first to validate the pertinence of a sustained vowel for perceptual voice analysis in French speakers and second to test the hypothesis that use of only the stabilized portion of the vowel would lead to underestimation of dysphonia. Voice samples were recorded in 60 dysphonic patients and 20 normal controls. Three different sample materials were obtained for each subject, i.e. connected speech, a sustained vowel including both the initial and stable portion (complete sustained vowel), and a sustained vowel including only the stable portion (stabilized sustained vowel). The jury was composed of 7 experienced listeners. Jury consistency was measured as a percentage of agreeing judgments for the same subject. No difference in consistency was observed using the three sample materials. Judgments on stabilized sustained vowels were confirmed as less severe than judgments on connected speech. Judgments on complete sustained vowels were comparable to judgments on connected speech.