一种基于声调规范模型的声调变换方法  

A Tone Transformation Method Based on Standard Tone Model

作　　者：薛健[1] 蔡莲红[2] 

机构地区：[1]清华大学继续教育学院,北京100084 [2]清华大学计算机科学技术系,北京100084

出　　处：《计算机工程与应用》2005年第10期40-43,85,共5页Computer Engineering and Applications

基　　金：国家863高技术研究发展计划项目(编号:2002AA117010);国家自然科学基金项目(编号:60275014)

摘　　要：该文利用固定点频率分析提取基音频率(F0),建立归一化线性多项式声调模型。参考男声、女声基音频率的分布和五度标调法,提出了一套汉语声调的规范模型,在此规范模型的基础上,实现了汉语语音声调变换。实际测听表明,经此模型变换的声音达到预期效果。In this article,method of Fixed Point Frequency Analysis was used to extract Fundamental Frequency(F0)as well as establish the normalized Liner Polynomial Tone Model.A set of standard tone models was proposed in light of the distribution of male voice,female voice fundamental frequency and five degrees of tone marked,by which way we have realized tone transformation of Chinese speech.By detecting of the hearing test,anticipated performance has been reached.

关 键 词：基音频率(F0) 声调规范模型 声调变换 

分 类 号：TP391.42[自动化与计算机技术—计算机应用技术]