Discriminating Neutral and Emotional Speech using Neural Networks

In this paper, we address the issue of speaker-specific emotion detection (neutral vs emotion) from speech signals with models for neutral speech as reference. As emotional speech is produced by the human speech production mechanism, the emotion information is expected to lie in the features of both excitation source and the vocal tract system. Linear Prediction residual is used as the excitation source component and Linear Prediction Coefficients as the vocal tract system component. A pitch synchronous analysis is performed. Separate Autoassociative Neural Network models are developed to capture the information specific to neutral speech, from the excitation and the vocal tract system components. Experimental results show that the excitation source carries more information than the vocal tract system. The accuracy neutral vs emotion classification using excitation source information is 91%, which is 8% higher than the accuracy obtained using vocal tract system information. The Berlin EMO-DB database is used in this study. It is observed that, the proposed emotion detection system provides an improvement of approximately 10% using excitation source features and 3% using vocal tract system features over the recently proposed emotion detection which uses the energy and pitch contour modeling with functional data analysis.