Direct and Indirect Measures of Glottal Volume Velocity

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): An ultimate goal of speech research is to develop a common theoretical framework that will link together physiology, acoustics, and speech perception. Such a unified approach is essential in explaining how tissue movement finally results in the perception of speech sounds. The proposed research evaluates a major assumption underlying the linear source-filter theory of speech production: that the glottal source and vocal tract are independent, and thus can be effectively separated using linear techniques. Because this same assumption underlies conventional inverse filtering (which is based on linear source-filter theory), results will also serve to validate this technique for estimating the glottal volume velocity. Estimates of the voice source from inverse filtering will be compared to the volume velocity calculated in a more direct but invasive manner. Because the volume velocity cannot ethically be measured directly in living humans, direct estimates will be generated in excised human laryngeal and in vivo canine models, by using anemometry to measure the exit jet particle velocity at the glottis and then multiplying this result by glottal area measures derived from high-speed imaging. Studies will examine a great range of normal and pathological phonatory configurations, and comparisons to inverse filtered data will lead to modifications in techniques to improve filter validity when necessary. By comparing volume velocity estimates from laryngeal modeling to those indirectly derived from inverse filtering the acoustic signal, the validity of inverse filtering for estimating the volume velocity will be established. Such studies will also illuminate the relationship between glottal dynamics and vocal acoustics, and thus will ultimately contribute to an enhanced basic understanding of the biomechanics of voice production and the practical limits of linear source filter theory in general. The primary hypothesis of this research is that the assumptions of linear source-filter theory hold for specific, limited regions of phonation. The goal of this research is to determine the regions over which these assumptions do not hold, and to quantify the interactions between source and filter across a wide variety of vocal configurations.