The onset of voice onset time: Measuring the emergent timing of laryngeal and oral gestures in early speech development

Abstract

Voicing contrasts in stop consonants are predominantly signaled by manipulation of voice onset time (VOT), the relative timing of oral and laryngeal gestures responsible for the release of the oral constriction and the onset of phonation. VOT is usually measured from the acoustic signal and defined by the time between the stop consonant burst and the following glottal period. Although this works well in adult speech, the same measure can be problematic when applied to child vocalizations. Infant productions of closant-vocant sequences perceived as consonant-vowel syllables by adults often do not exhibit full closure and are voiced throughout. Traditional measures of VOT cannot then be employed, even though evidence from the acoustic signal can still be used to deduce the presence and timing of precursors of the oral and laryngeal gestures from which adult voicing contrasts later emerge. This study illustrates these issues using data derived from home audio recordings made monthly from birth to three years as part of a large-scale, longitudinal study of infant vocal development. New methods for measuring VOT are proposed, using the relative amplitude, frequency and phase of the harmonics in the acoustic signal to infer the timing of underlying articulatory gestures.