Thesis (Ph.D.) - Cairo University - Faculty of Medicine - Department of E.N.T

Acoustic analysis including both spectral and cepstral analyses is an objective and easy assessment tool for diagnosis and follow up of dysphonia. Objective: the aim is to investigate the cepstral voice analysis as a valid method for assessment of dysphonia and comparing it to the traditional spectral voice analysis in a sample of Egyptian population. Subjects and Methods: the study was carried out by applying a voice assessment protocol used in Phoniatric unit of Cairo University on 140 subjects who were divided into 2 groups, dysphonia group includes 70 subjects and control group includes 70 subjects. The voice samples in the form of sustained vowel /a/ phonation and 6 different contexts of continuous speech of all subjects were recorded and assessed using auditory perceptual analysis, spectral analysis and cepstral analysis. Results: The present study revealed statistical significant difference for both traditional (time based) and cepstral analyses between cases and control groups and among different grades of dysphonia during sustained vowel and continuous speech conditions apart from speech context 1. Jitter analysis was found to have the best predictive ability during sustained phonation with sensitivity of 92.86% and a specificity of 82.86%, while measures of CPPs analysis were found to have the best predictive ability during continuous speech with sensitivity of 88.57% and a specificity of 77.14%

Issued also as CD