Investigation into Voice Source of Monozygotic Twins using Formant based Inverse Filtering

Abstract

Studying glottal flow gives potential benefit in many disciplines. Several methods have been developed for the estimation of the glottal flow. Glottal flow can be estimated from microphone pressure signal, or Inverse filtered signal. Inverse filtering (IF) has been used widely for the understanding of phonation type, intensity, voice quality, emotions and vocal loading. However there is a dearth of information about the effectiveness or reliability of IF techniques especially in the evaluation of individuals having similar voice and speech characteristics like monozygotic twins. Also, voice source through inverse filtering has not been investigated in twins so far. In this context, the present study investigated similarity of voice source in monozygotic twins using inverse filtering and the consistency of inverse filtered parameters. Two groups of females participated in the study. Group I had 6 monozygotic twins and Group II had 6, age and gender matched unrelated pairs. None of them had any voice disorders. Subjects Phonated vowel /a / three times at least for 5 seconds in comfortable pitch and loudness. Samples were audio-recorded at a sampling rate of 48 kHz and phase linear recording. Samples were inverse filtered using Vag_physio module of VAGHMI software in formant based method. Results showed that IF parameters were reliable over the repeated trials in all individuals. Also, ANOVA showed no significant difference between groups on voice source characteristics. The open quotient (OQ) and speed quotient (SQ) was significantly different across groups. However further investigation on twin pairs selection based on perceptual similarity and confirmed genetic analysis is warranted.