Computational Estimates of Pitch Height and Pitch Strength in Dysphonic Voices

Voice disorders affect millions of people worldwide and result in significant negative impacts on their social, professional, and health-related qualities of life. Acoustic analysis of voice is an indispensable part of clinical assessment of dysphonia and evaluation of treatment outcomes. However, the accuracy and validity of most conventional analysis routines is limited by the degree to which the voice is periodic. For many dysphonic voices with severe aperiodicity, acoustic measures are often invalid and at best, are error-prone. The current study overcomes this periodicity limitation through auditory-based measurement methods and automated computational estimates of pitch height (scale of low to high) and strength (scale of weak to strong).