Respiratory constraints on speech production at prosodic boundaries

Abstract

This research characterizes the respiratory system dynamics at the initiation and termination of utterances and determines correlations of physiological measures with acoustic cues for these prosodic boundaries. The analysis includes boundaries within a breath as well as boundaries that are aligned with the initiation and termination of exhalation. Simultaneous recordings of the acoustic signal, airflow, esophageal pressure and lung volume were collected during read isolated utterances and short paragraphs. These measures were used to derive estimates of recoil forces of the chest wall, net muscular forces, and the area of the airway constriction. Data are presented from four subjects (two men, two women), all native speakers of American English. Perceptual ratings for initial and final prominent syllables and the locations of pauses within the utterance were also collected. For speech boundaries th.i.t are aligned with breath boundaries, utterance initiation occurs during a rapid transition in muscular effort. Sound begins as soon as conditions permit and these conditions consistently occur during net inspiratory muscular force. Alveolar pressure reaches an initial peak (PpI) that is, in most cases, correlated to the relaxation characteristic of the chest wall. The timing of Pp1 generally coincides with a prominent syllable if that syllable is the first or second syllable in the utterance and precedes later prominences. Pressure at phonation onset is, on average, near 0.3PpI for utterances initiated with a voiced sonorant and is near 0. 8Pp1 for utterances initiated with a voiceless fricative. Phonation termination results from an approximately 3-fold increase in glottal area and a J-3 cm H20 fall in pressure. Irregular fundamental frequency (FO) at the end of voicing, in many cases, does not fit the classical definition of glottalization. Instead, voicing terminates with increasing glottal area, and FO becomes irregular during the increase. In some cases, regular FO resumes as glottal area continues to increase. Distinct respiratory gestures are made at pauses within a breath. The pressure is reduced by 2-3 cm H20, on average, during a period of relatively little volume change. The findings in this research show that the role of the respiratory system in speech production goes beyond a more traditional view of this role as one of simply providing a relatively constant driving pressure during speech.