Using prosody database in Chinese speech synthesis

As the difficulty of revealing the relationship between text and prosody, a corpus based approach is used in prosody generation in the research. A prosody database is built as templates for prosody generation. The general idea is that we are trying to get the prosodic information from real speech examples. We first analyze given Chinese text, and form a linguistic feature vector, which describes the phonetic and lexicon characteristics of the text. Then we search the database to find the best match of the vector, which is a similar occurrence of the text. The prosody parameters of the retrieved example will be the guideline of the prosody we are going to generate.