Spectral Analysis of Hypernasality

A Lee, V Ciocca, T Whitehill


The aim of this study was to assess the effectiveness of one-third-octave analysis for evaluating hypernasality. Subjects included 20 speakers with hypernasality and 20 normal controls. The vowels //, /i/, // and /u/ were segmented from 18 Cantonese single words produced in two consonant contexts, oral and nasal. The results showed statistically significant differences between the experimental and control groups for vowels /i/ and // spoken in oral context (e.g. /bit/) and vowel // in nasal context (e.g. /mN/). The one-third-octave bands that separated the two speaker groups varied among vowels but the band with a center frequency of 800 Hz consistently distinguished the two speaker groups in the above three vowels. Several other one-third-octave bands revealed significant differences between the two groups for the vowel /i/ produced in oral consonant context. Bands in both oral and nasal consonant contexts could distinguish between the two speaker groups for the vowel //.