A Corpus-Based Prosodic Modeling Method for Mandarin and Min-Nan Text-to-Speech Conversions

This talk gives an introduction to a recurrent neural network (RNN) based prosody synthesis method for both Mandarin and Min-Nan text-tospeech (TTS) conversions. The method uses a fourlayer RNN to model the dependency of output prosodic information and input linguistic information. Main advantages of the method are the capability of learning many human’s prosody pronunciation rules automatically and the relatively short time of system development. Two variations of the baseline RNN prosody synthesis method are also discussed. One uses an additional fuzzy-neural network to infer some fuzzy rules of affections from high-level linguistic features for assisting in the RNN prosody generation. The other uses additional statistical models of prosodic parameter to remove some affecting factors of linguistic features for reducing the load of the RNN.