Control of laryngeal dynamics in speech and song

• Grant number: 22/07143-0
• Support Opportunities: Scholarships abroad - Research
• Start date: July 01, 2023
• End date: February 01, 2024
• Field of knowledge: Linguistics, Literature and Arts - Linguistics
• Principal Investigator: Beatriz Raposo de Medeiros
• Grantee: Beatriz Raposo de Medeiros
• Host Investigator: Khalil Iskarous
• Host Institution: Faculdade de Filosofia, Letras e Ciências Humanas (FFLCH). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: University of Southern California (USC), United States

Abstract

In this research project we aim to build a dynamical statistical model of the human voice system that explains how the vocal folds oscillatory movements can be both extremely stable, as the nonlinear dynamics of flow and vibration require absolutely stable dynamics, but also be extremely flexible as to allow for very fast and complex changes in vocal dynamics as observed in tonal and intonational behaviors in speech, as well as many singing behaviors. We will use state-of-the art algorithms from signal processing and statistics, as well as recent developments in theories of speech motor control to establish a theory of the control of the voice mechanism that simultaneously addresses stability and flexibility issues. Previous theories of vocal fold vibration in the speech and singing literatures address either stability or flexibility, while the innovation of the work on this project will be to address both. Databases of speech acoustics in English and Portuguese speech and song will be used to develop a dynamical model using Bayesian induction techniques that are flexible enough to enable the estimation of model parameters, as well as speaker, register, and gender variation. The answers to be reached will contribute to a better comprehension of the facts responsible for the phonatory function of the larynx and will lead us to understand and explain human phonation. Methods will include three main steps: (1) literature review on mathematical models of vocal fold vibration (2) literature review on linguistic and singing control of the laryngeal motor system and neurophysiology; (3) development of the larynx model description based on limit cycles and bifurcation theory, and Bayesian statistics. (AU)