摘要
近年来，在汉语语音合成的研究中，如何提高合成语音的自然度越来越引起人们的关注．其中的关键问题是设计一个能灵活有效地调节汉语韵律特征的语音合成器．进行韵律调节的基本原理是基于源／滤波器理论，采用同态处理的方法，对汉语文语转换系统进行研究，并对韵律调节部分进行了分析和实验．实验结果显示，这种方法要胜于ＰＳＯＬＡ方法．下一步，一块基于ＤＳＰ芯片“ＴＭＳ３２０Ｃ５４ｘ”的ＰＣ卡将用来实现上述目标，从而期望建立一个实时的语音合成器。
In the recent study of Chinese speech synthesis,more attention is
focused on improving the naturalness of the synthetic speech.The key point is how to design a
speech synthesizer with flexible prosodic adjustment. The principle of prosodic adjustment is
based on source/filter theory,and the homomorphic analysis method is used in present paper.A
Chinese test to speech system is used for experiments.The result of the experiments shows
that this method is better than PSOLAs.As the next step, a PC card based on a DSP chip
TMS320C54X will be used to implement the computing task. It can be expected that a real time
synthesizer,which will have wide applications in speech synthesis, will be demonstrated.
出处
《复旦学报（自然科学版）》
CAS
CSCD
北大核心
1999年第3期325-330,共6页
Journal of Fudan University：Natural Science
基金
国际协作项目
日本NTTDATA公司资助课题
