Acoustic features mediating height estimation from human speech

Abstract

The current experiment was aimed at providing the first direct evidence regarding the role of subglottal resonances in height discrimination. Past research has investigated whether or not listeners are able to discriminate which of two talkers is taller (Rendall et al., 2007), but has not established what parameters of the acoustic speech signal listeners use to distinguish speaker height. In the current study, we examined the role of subglottal resonances in height discrimination. Subglottal resonances generally refer to resonances of the lower airways starting in the lungs and terminating at the glottis. Subglottal resonances would be a good candidate for use in height estimation because they remain relatively stable across vowels and are relatively unaffected by other factors influencing the supra-laryngeal vocal tract; instead they are influenced nearly exclusively by the lower airway acoustic properties. Listeners participated in two tasks, the first task was a two-alternative forced choice (2AFC) height discrimination test, in which listeners heard sentences produced by two talkers of the same gender and were asked to determine which of the two was the tallest. The second task involved listening to five talkers (all same gender) sentences, and then ranking those individuals from tallest to shortest. Findings indicate that listeners are able to discriminate and rank speakers heights better than chance, and the role of subglottal resonances will be reported.