Wavelet transform in the time-frequency analysis of patients with articulation disorders

Abstract

Spectral analysis of the human voice is a frequently used digital analysis method in the diagnosis, the planning and follow-up of the treatment of speech disorders. In the classical spectral analysis method, the principals of Joseph Fourier are used. This is called "Fourier Transform" and it accepts that all signals are formed of the synthesis of many sinusoïdal formed signals. In recent years a new transform method called "wavelet transform" accepts the complex signals formed of small signal particles called "wavelets" and it is considered that this transform will solve the documented problems of the "Fourier Transform". By using the appropriate wavelet, this transform can be used as an alternative to the Fourier transform. In this study, the patients with an articulation disorder of the "s" sound were evaluated before and after the phoniatric reeducation by using both the transform methods, and the results obtained are discussed.