Segmental durations in connected-speech signals: Current results

Abstract

Two types of analyses have been performed on the measured durations of recordings produced by six talkers reading two scripts of approximately 300 words each. The texts, the combined visual–auditory marking technique, and preliminary results were reported earlier by Crystal and House [J. Acoust. Soc. Am. 72, 705–716 (1982)]. The average durations and standard deviations of various classes of speech sounds, as well as individual speech sounds, have been determined and segmental measurements are compared to earlier data and to various pertinent published reports. The histograms of the measured durations of various sounds and categories have been fitted with distributions which are, equivalently, the exit-probability sequence for a Markov chain or the impulse response of an IIR digital-filter network.

Informativity affects consonant duration and deletion rates

“…They contrasted the average properties of different segment types in several environments, discussing phenomena such as the effect of stress on segments, the effect of voicing on preceding vowels, and more. While Crystal and House (1988a) report their findings for individual segments as well, the number of tokens per segment is relatively small (e.g. only 224 /p/s in all environments in Crystal and House [1988a]), making it difficult to control for multiple factors.…”

“…While Crystal and House (1988a) report their findings for individual segments as well, the number of tokens per segment is relatively small (e.g. only 224 /p/s in all environments in Crystal and House [1988a]), making it difficult to control for multiple factors. Lavoie (2001) contrasted the articulation of consonants in American English and Mexican Spanish in several environments, investigating several properties such as manner of articulation, duration, and intensity.…”

“…Previous research has shown that many phonetic factors affect the surface realization of phonemes. For instance, voiced stops tend to have shorter duration than voiceless stops (Crystal et al 1988a;Lavoie 2001; among others), but their actual realization differs across languages. Comparative phonetic studies (Lavoie 2001;Torreira and Ernestus 2011) found systematic differences between the realization of otherwise similar segments in different languages.…”

“…Lavoie (2001) contrasted the articulation of consonants in American English and Mexican Spanish in several environments, investigating several properties such as manner of articulation, duration, and intensity. As Crystal and House (1988a) did, Lavoie (2001) focused on summary statistics, and the number of tokens in her sample makes it impossible to control for multiple factors at once.…”

“…Crystal and House (1982;Crystal and House 1988a, Crystal and House 1988b, Crystal and House 1990 investigated the duration and variability of English consonants and vowels in connected speech. They contrasted the average properties of different segment types in several environments, discussing phenomena such as the effect of stress on segments, the effect of voicing on preceding vowels, and more.…”