Spectrum analysis of vocalization application for voice pathology detection

Abstract

In this paper we present an approach to identify pathological disorders of larynx using spectrum analysis of speech samples. The speech signal is subjected to spectrum analysis, and the harmonize coefficient is introduced to identify disorders. Distinctions between norm and pathology are observed concerning spectral component of a high-frequency and low-frequency range. Examples of application of this system for inspection of people with diseases of the voice device and healthy people, including -professional singers are presented.