The Pisan Vowel System of Read and Semispontaneous Speech. An Exploratory Contribution

Speech style is an important factor with respect to vowel reduction, and the availability of larger and more realistic samples of data allows the investigators to focus on connected speech processes. The main goal of the analysis reported on in this paper is to investigate the dynamics of vowel systems under speech style changes. Utterances of stressed vowels both in semispontaneous- and read speech were analysed; the data came from the speech of two young subjects from Pisa, recorded in the AVIP (Archivio delle Varietà di Italiano Parlato) project. Works with a similar experimental design are available for Castilian Spanish (Poch-Olivé & Harmegnies 1992, Harmegnies & Poch-Olivé 1992), Belgian French (Poch-Olivé & Harmegnies 1992), Portuguese (Delplancq et al. 1995-96), English (Deterding 1997) and Italian (Giannini & Pettorino 1993; Poch-Olivé & Harmegnies 1995). Duration, F1, F2, F3 and F0 of nearly 700 vowel utterances were measured. The data collected were presented in terms of F1/F2 diagrams on a Hz scale. As for duration, F1, F2, F3 and F0 sets of ANOVAs were run to test the effects of the variable ‘speech style’ for the seven vowels /i e E a o u/. In order to estimate how peripheral the vowels were, the Euclidean distance of each vowel from the centroid of all vowels in both speaking styles were calculated, and statistical results of the values thus computed for each vowel and speech style were presented. The acoustic size of the vowel systems in both speech conditions was calculated: the space was largest in the read speech condition and smallest in the quasi-spontaneous speech condition. The formant and duration variabilities were further observed: a lowered differenciation of the sounds in the quasi-spontaneous speech was therefore noticed. The formants frequencies were finally processed by means of discriminant analysis, which confirmed that vowels in the F1/F2 space were more differenciated in laboratory- than in semispontaneous speech.