The Influence of Speaking Rate on Vowel Formant Track Shape as Modeled by Legendre Polynomials

Speaking rate and vowel duration are generally thought to affect the dynamic structure of vowel formant tracks. This idea was tested by letting a single professional speaker read a long text at two different speaking rates, fast and normal. The extent to which the shape of the first and second formant tracks of 8 Dutch vowels varied under the two different speaking rate conditions was investigated. A total of 549 pairs of vowel realizations from various contexts were selected for analysis. Legendre polynomial functions were used to model and quantify the shape of normalized formant tracks. No differences in normalized formant track shapes were found that could be attributed to differences in speaking rate. But a higher FI frequency in fast rate speech relative to normal rate speech was found that can be explained as the result of a uniform change in frequency. These results indicate a much more active adaptation to speaking rate than implied by the target undershoot model. Within each speaking rate. there was only evidence of a weak leveling off of the FI tracks of the open vowels /£, a, a / with shorter durations. These same conclusions were reached when sentence stress was taken into consideration and when vowel realizations from a more uniform, alveolar-vowel-alveolar, context were examined separately. In the alveolar context, a small rise in F2 of the vowel /o/ might indicate more coarticulation in fast rate speech.