Multidimensional acoustic analysis for voice quality assessment based on the GRBAS scale

摘要

Despite the fact that perceptual evaluation of voice quality is considered as a gold standard for examining normal and pathological voice quality, the considerably high inter- and intralisteners variability still cannot be neglected. This is the result of a number of confounding factors such as listeners' perceptual bias, listeners' experience and type of rating scale being used. Currently, automatic objective assessment provides a very useful tool for diagnosis of pathological voices. Acoustic analysis can be a useful complementary tool for determining severity of dysphania. The present study aimed to develop a complementary automatic assessment system for voice quality by using multidimensional acoustical measures based on the well-known GRBAS scale. A total of 65 dimensionality measures including Mel-frequency Cepstral Coefficients, Glottal-to-Noise Excitation Ratio, Vocal Fold Excitation Ratio were constituted a set of features. Additionally, to reduce redundancy of providing features, three different feature extraction techniques were applied. The multiclass classification was done by means of RBF kernel-SVM. The classification results were moderately correlated with GRBAS ratings of severity, with the best accuracy around 70%. This suggests that such multidimensional acoustic analysis can be an appropriate assessment tool in determining the presence and severity of voice disorders.