Abstract: To meet the actual needs of Uyghur dialect speech synthesis, this paper gives priority to Kashgar dialect. Based on the comparison with text corpus, the paper comparatively analyzes the phoneme correspondence between the standard Uyghur language and Kashgar dialect, observes the changes like phoneme insertion, deletion and substitution phoneme, collects the corresponding rules for these changes, and constructs the finite state transducers for those conversion rules. Experimental results show that the automatic conversion algorithm based on rule has good performances.

Key words: Uyghur, Kashghar dialect, phoneme, phoneme change, dialect conversion

摘要： 从喀什方言维吾尔语音合成的实际需求出发，在对照文本语料库的基础上，对比分析从标准维吾尔语到喀什方言的音位对应关系，根据增音、减音、换音等变位现象，总结相应的音位变化规则，构造基于有限状态自动机的转换模型及算法。实验结果表明，基于规则的自动转换算法具有良好的运行效果。

关键词: 维吾尔语, 喀什方言, 音位, 音位变化, 方言转换

CLC Number: 

• TP391