摘要： 根据语音合成与识别等语音应用研究的需求，从文本分析模块入手，选择“维吾尔语语音声学参数库”中包含清塞音p、t、k、q的单音节及多音节词，对它们的声学参数采取统计分析方法，归纳其嗓音起始时间、共振峰、音强和时长分布模式。同时，为提高语音合成的自然度，探讨嗓音起始时间对清塞音声学特征的决定性作用问题。

关键词: 清塞音, 嗓音起始时间, 共振峰, 时长, 音强

Abstract: According to the needs of speech processing applications such as how to improve the naturalness of speech synthesis system, this paper presents for the first time the systematic empirical acoustical features analysis results of the tenuis consonants in Uyghur language based on the “Uyghur Acoustical feature Database”, which contains acoustical measurements of tenuis consonants contained in words such as p, t, k, q, which are recorded by 1 male and 1 female speakers separately, and finds the basics for the studying the prosody in words and sentences of Uyghur language. The aims and research results of this paper is to get a high value for further studying the Uyghur natural language processing techniques.

Key words: tenuis consonant, voice onset time, formant, time length, intensity

中图分类号: 

• N945