摘要
为了解决中文语音合成芯片构成的语音系统中播放语音速率不可调节的缺点,提出了先用特殊字符对输入的文本进行标注,把文本分成了不同类型的信息帧,然后将信息帧传输给单片机,系统自动判断信息帧的类型,根据不同的信息帧设置不同的延迟时间,从而达到调节播放语音速率的目的.通过实验测试结果表明,这样不仅实现了调节播放语音速率的目的,同时也提高了播放语音的可懂度和自然度.该系统可以利用在各种嵌入式语音合成的系统中.
A method for solving the problem that Chinese speech synthesis chip cannot adjust the playback rate is proposed in this paper.First,special characters are used to mark the input text and divide it into the different types of information frames,and then the information frames are transmitted to a single chip microcomputer.Automatically judging the types of the information frames,the system sets different delay time to achieve the desired adjusting playback speed.An experiment showed that this system can produce speech with high intelligibility and naturalness and is recommended to be applied in embedded Chinese speech synthesis systems.
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第3期153-157,共5页
Journal of Southwest University(Natural Science Edition)
关键词
语音合成
信息帧
文语转换
speech synthesis
information frame
TTS（text-to-speech）
