Acoustic Analysis and Modeling of Pathological Voices

Опубликовано 6 сентября 2016, 16:45

Disorders in voice are common and presently have been increasing due to unhealthy habits and vocal abuse. As well, there are a great range of diseases which cause modification of the speech, for example which belong to a class of neuro-degenerative diseases. It is well known that most of the vocal and voice diseases cause changes in the acoustic voice signal. Acoustic analysis of the speech signal may be used as complementary tool in laryngoscopy to diagnose this sort of disease. Speech processing is particularly interesting due to the simplicity and the noninvasive nature of the measurement procedure for the achievement of the acoustic voice analysis. In the presentation, principles and previous studies will be shown to specify and model the most significant acoustic characteristic of disordered voices as basis for the development of effective methods on the purpose of discriminating between normal and pathological voices or for predicting perceived voice quality or the severity of voice disorder. The acoustic modeling of pathological voice is based on a parametric representation of speech, as well a nonparametric approach, and a statistical approach for classifier design.