AUDITIVE LEARNING BASED CHINESE FO PREDICTION

摘要

The paper describes a new F0 model based on auditive learning (AL) method. Being focused on the notion of prosody templates, we confirmed that FO patterns for a syllable can be extracted from various anamorphosis of FO contours in spontaneous speech. It is much suitable to use FO templates selection method for Chinese F0 prediction with prosody cost function (PCF). Furthermore, an AL method is used to adjust the weight of PCF dynamically in application. Unlike other methods, the approach may give feedback as to exactly what are the crucial parameters determining the successful choice of patterns. The paper also analyzes the error distribution of the FO predicting results. Both smoothing testing and FO range testing show that the synthesis results are much closed to human being.