Influences of various factors upon parameters of the command–response model for fundamental frequency contour generation

Abstract

The command–response model by Fujisaki and his co-workers formulates the generation process of the fundamental frequency contour (henceforth F0 contour) in terms of a set of input commands carrying linguistic and paralinguistic information and the mechanisms that respond to these commands. The parameters of the mechanisms are time constant of the phrase control mechanism (α), that of the accent control mechanism (β), and the baseline frequency (Fb). The present study investigates the influences of speech rate, speaking style, and individual difference on these parameters as well as their variability. The speech material consisted of recordings of a short story by six native speakers of Japanese at three speech rates (slow, normal, and fast) and in two speaking styles (reading and conversational). The parameters were extracted by the method of analysis-by-synthesis, and the results were analyzed statistically. The analysis indicated that utterance-to-utterance variations are quite small in all three parameters for a given rate, style, and speaker. Among the three parameters, only α showed a small but systematic tendency to increase with the speech rate, while differences in speaking style did not affect these parameters. Individual differences were quite small in α and β, while Fb varied from speaker to speaker.