Modelling Speech Melody

The search for an appropriate scale for measuring fundamental frequency has been one part of a systematic attempt, in particular by researchers from the Netherlands (’t Hart et al., 1990) (“the Dutch school”), to develop a model of the way in which pitch is perceived. This was done by stylising raw fundamental frequency patterns as a sequence of straight lines, such that when the stylised frequency is used to resynthesise the utterance, the result is judged to be perceptually equivalent to the original intonation pattern. For more recent work extending this approach to analysing intonation see Mertens (2004, 2022), which draws on earlier work on tonal perception by House (1990). Mertens’ tonal perception model makes the assumption, contrary to what I will argue below, that the perception of a pitch contour depends on the prior segmentation of the utterance into syllables, so that a pitch contour on a polysyllabic word will be perceived in a fundamentally different way from the ‘same’ contour on a single syllable.