A phonetic model of the human pharynx
Date
1998
Authors
Heap, Lynn Marie
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Esling ( 1996) has proposed that the range of speech sounds that are articulated behind the velum can be organized according to one place of articulation with four possible manners of articulation. This thesis provides acoustic and video-endoscopic evidence for the re-classification of post-velar sounds by describing pharyngeal speech sounds with criteria used in existing phonetlc taxonomies and theories of linguistics. According to the International Phonetic Alphabet (IPA) taxonomy, there are sounds that have been identified as pharyngeal as well as sounds classified as epiglottal. Esling has proposed that the epiglottal sounds can be re-classified as varying manners of articulation that are produced in the pharynx.
This thesis uses acoustic and video-endoscopic evidence to show that there is one place in the pharynx that has major construction capabilities, and that the larynx as a whole is able to move independently beneath the sphincter that is created by the aryepiglottic folds and the epiglottis. It is shown that the larynx can rise and lower vertically beneath this pharyngeal construction, which contributes to the auditory confusion of post-velar sounds. Acoustic features that are associated with raised larynx and pharyngealized voices in Laver's taxonomy of Voice Quality Settings are used to provide support for larynx height as a distinctive feature in pharyngeal sounds.
Along with the correlation of acoustical information with linguistic taxonomies, other acoustical correlates including pitch, shimmer, jitter and harmonic-to-noise ratio are described for cardinal pharyngeal sounds as well as for natural language data that have pharyngeal sounds. Based on the description of acoustical effects that identify pharyngeal sounds there needs to be a modification of the traditional acoustical model of speech production.
The existing model of speech production, the source-filter theory, requires modification, since pharyngeal construction is shown to have an effect on the source aspect of the model by decreasing vocal fold vibration. The ability of the aryepiglottic folds to vibrate simultaneously is shown to provide another source of sound that can occur in conjunction with vocal fold vibration The filtering acoustic effects of pharyngeal construction and changes m larynx height have only been generally described in the past. The closer acoustical examination of cardinal pharyngeal sounds m this thesis suggests that it is energy in the upper frequency range that distinguishes between changes m the back part of the vocal tract.
Also developed in this thesis is a technique for measuring video images. By normalizing video images to a common reference in the vocal tract, it is possible to compare pictures between subjects and recording sessions. The results of applying the video measurement technique to images of cardinal pharyngeal sounds suggest that this technique will be more successful in describing voice quality settings when applied to a sequence of video pictures.
This thesis, in describing pharyngeal speech segments, has provided a range of acoustical measurements that are found to be salient in describing pharyngeal sounds and manners of articulations The result of this broad acoustical analysis has shown that the present source-filter theory of speech production needs to be modified in order to accommodate the observed construction capabilities in the pharynx and vertical movements of the larynx. The video measurement technique developed m this thesis allows video pictures to be compared so that quantitative details between subjects and recording sessions can be stated.