Effects of Breathy Voice Source on Ratings of Hypernasality

Abstract

To investigate the effects of breathy voice sources on ratings of hypernasality using synthesized speech. Speech samples were obtained from children with cleft palates who demonstrated varying degrees of hypernasality and from a child with a voice disorder. Sources with 6 degrees of breathiness were created: a breathy source and five synthesized sources with lowered harmonics-to-noise ratio (HNR) values by the addition of impulses. These sources and each original (clear) source were combined with three kinds of filters: mild, moderate, and severely hypernasal. Consequently, 21 ([6 + 1] x 3) stimuli for each vowel (/a/ and /i/) were obtained for ratings. Thirteen speech pathologists with academic training and various clinical experiences with cleft palate speech rated hypernasality of the stimuli on a 5-point scale. Ratings of hypernasality for breathy and clear stimuli were analyzed using a repeated measures analysis of variance. The effects of breathy source on ratings of hypernasality were significant for the following filters: mild hypernasal /a/, severe hypernasal /a/, mild hypernasal /i/, and moderate hypernasal /i/. A post-hoc comparison test demonstrated that the more breathy sources (BH0 or BH2) generally increased the hypernasality score for mild hypernasal filters and decreased it for moderate and severe hypernasal filters. The less breathy sources (BH3, BH4, and BH5) hardly affected the ratings. The effects of breathiness on ratings of hypernasality seem to moderate rather than to mask perceived hypernasality. That is, breathiness raises slight hypernasality, whereas it reduces severe hypernasality.