Emotion Recognition and Conversion for Mandarin Speech

In this study, some research activities on expressive speech recognition and conversion will be introduced. A database consisting of five kinds of speech emotions (i.e. happiness, sadness, surprise, anger and neutral) is used. Not only those traditional features such as mfcc, plp, and pitch are studied, but also a new feature extraction method based on fisher´s F-Ratio is proposed and reported. In our experiments, various combinations of these features, including their high order features are applied using GMM modeling for Mandarin expressive speech recognition. Also we presented some results from emotional speech conversion with a pitch target model.