基于最小合成单元的维吾尔音库设计  

Smallest Speech Unit Corpus Design for Waveform Concatenation based Uyghur TTS

作　　者：卡斯木江.卡迪尔 古丽娜尔.艾力 艾斯卡尔.艾木都拉[1] 

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046

出　　处：《通信技术》2012年第4期83-85,共3页Communications Technology

基　　金：新世纪优秀人才支持计划资助;国家自然科学基金资助项目(批准号:61065005;61062008)

摘　　要：为了实现容量小、可懂度和自然度较好的波形拼接式维吾尔语语音合成系统,结合维吾尔语的特点,完成了文本设计、录音、语音标注、以及音库建设过程。其中,把音节作为基本的合成单元建立了音节语音库;为了弥补音库中不存在音节的合成问题,又建立了音素作为合成单元的音素语音库。实验结果表明,以音节和音素为最小合成单元的波形拼接式维吾尔语语音合成系统除了音库容量相对少等特点之外,还具有良好的可懂度。In order to achieve waveform concatenation-based Uyghur speech synthesis system,with smaller capacity,better intelligibility and naturalness,and in combination with the characteristics of the Uyghur language,the processes of text design,recording,labeling and corpus construction are done,and with syllable as the basic synthesis unit,the syllable corpus is established.For non-existence of syllable synthesis in the corpus,a phoneme corpus with phoneme as the synthesis unit is also established.Experimental results show that the proposed waveform concatenation-based Uyghur speech synthesis system with syllable and phoneme as the smallest synthesis unit,has relatively small corpus,and its synthesized speech is of fairly good intelligibility.

关 键 词：维吾尔语 语音合成 语料库 特征提取 

分 类 号：TN912[电子电信—通信与信息系统]