Spectro-temporal description of dipthongs in F1-F2-F3 space

Abstract The prevailing approach to the acoustic-phonetic description oescription of the diphthong is based (1) on the two lowest, vocal-tract resonance (or formant) frequencies ( F 1 and F 2 ) considered either individually or jointly in the F 1 − F 2 plane, and (2) on a very sparse representation of the temporal course of these frequencies. While this time-honoured approach has been particularly useful for characterising the initial and final vowels of the diphthong, there appears to be very little progress beyond the F 2 − F 2 plane, as a parametric framework for elucidating the dynamic nature of the vowel-to-vowel transition. By contrast, a more accurately spectro-temporal description of a subset of the Australian English diphthongs ( ) is obtained in this work by considering a detailed, temporal representation of the three lowest formant-frequencies ( F 1 , F 2 and F 3 ). In particular, certain nonlinearity features of the densely-sampled contour of the F 3 are highlighted, which appear to have hitherto been either unknown or considered inconsequential to the specification of the diphthong. This finding is shown to contribute a new, three-dimensional ( F 1 − F 2 − F 3 ) perspective on the acoustic characteristics of the vocalic transition of the diphthong.