Objective Acoustic Analysis and Comparison of Normal and Abnormal Voices

Abstract

Introduction: Acoustic analysis is commonly used to diagnose, document and treat voice disorders. Type I and Type II voices which are nearly periodic can be easily assessed with computerised acoustic analysers. Widely used voice parameters like Jitter, Shimmer and Harmonic Noise Ratio (HNR) are indices of voice pathology and indicate different diseases and help as an outcome measure. Aim: To study whether objective acoustic analysis is able to differentiate between abnormal and normal voices. Materials and Methods: In the present study, the patients were made to phonate sustained vowel /a/ and the voice was recorded to analyse parameters like Jitter (ddp), Shimmer (dda), HNR and Median pitch using the acoustic software, PRAAT. These parameters were compared with the values of the institute’s own normal voice database. They were analysed for Mean, Median, standard error of mean, and KolmogorovSmironov values. Results: The range of abnormal voice parameters like Jitter (ddp), shimmer (dda), median pitch and HNR measures were different from normal voices. The difference was significant in jitter (p-value of 0.026) in males, in Shimmer (p-value of 0.035) in females. HNR did not show any significance. Conclusion: The traditional methods of perturbation measures like Jitter, Shimmer, Median pitch and HNR can help the clinicians for characterisation of voice into either normal or abnormal voices. But the comparison needs a local or regional normal voice database.

Keywords

• hoarseness of voice
• voice assessment
• voice disorder