一种数据驱动的汉语成句语音韵律特征产生模型的研究

针对汉语发音特点，基于对大量自然汉语语句基频轮廓数据的统计和分析，提出一种用于数据驱动生成汉语韵律特征的数学模型。该模型以基频参数为主，辅以时长和增益参数，能表现汉语的语气、短语节奏、韵律词声调及轻重音多层韵律信忠，各层参数可按语言知识分类训练和标注。给出了模型的各种归一化“调素”函数和变调规则。仿真实验表明了该模型的有效性。

Study of data-driven hierarchical prosody generation model for Chinese sentence utterance

Aim for Chinese speech, a large number of typical utterance pitch contours are extracted and analyzed, a data-driven prosody generation model for Chinese TTS is presented, which is mainly characterized on pitch parameter, combined with the duration and the gain. It incorporates hierarchical Chinese prosody messages, such as the sentence mood, phrase pace, tone and accent. A set of practical normalization multi-tone pattern functions and the sandhi rules are provided. Emulation tests show that the model and rules effectively represent the relationship between the prosodic features and the multi-layer linguistic information.