摘要
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。
This paper describes prosodic control rules for Wu-dialect Chinese text-to-speech system. It is different between Mandarin Chinese and Wu-dialect Chinese not only in tone numbers, types and tone transition in continuous speech, but also in length, pause and intonation. Therefore, the special prosodic control models have to be proposed to control the speech synthesis for a Wu-dialect Chinese text-to-speech system. In this paper, several prosodic control rules for tone control, length control, strong ness control etc.were proposed after a short introduction of prosodic characteristics of Wu-dialect Chinese, the experiments of speech synthesis with prosodic control and evaluation of them were presented then.
出处
《通信技术》
2002年第10X期1-3,9,共4页
Communications Technology
关键词
韵律控制
文语转换
语音合成
吴语
prosodic control, text-to-speech conversion, speechsynthesis, Wu-dialect Chinese