论文部分内容阅读
吴语不同于汉语普通话,无论是声调类型与特征、协同发音的变调方式,还是语句的语调与幅度变化、单词重音位置与时长,都有其自身的规律。因此,在吴语文语转换中有必要建立相应的韵律控制模型来控制语音的合成输出。这里首先介绍吴语的语音韵律特征,然后介绍语音合成中具体的声调控制、变调处理以及时长与幅度调整处理方法,对语调的处理也作了简单介绍,并给出了运用韵律控制合成语音的实验评价结果。
Wu dialect differs from Mandarin Chinese both in terms of tone type and characteristics, co-articulation tone variation, intonation and amplitude of sentences, and location and duration of word stress. Therefore, it is necessary to establish the corresponding prosodic control model to control the synthesized output of speech in Wu language. Here we introduce the speech prosody features of Wu language firstly, and then introduce the specific tone control, pitch change processing and the method of duration and amplitude adjustment in speech synthesis. We also give a brief introduction to the processing of intonation, and give the experiment of using prosody to control synthesized speech Evaluation results.