中文摘要：

方言语音的转换是人机交互领域的一个重要研究课题。为实现普通话到西安话的转换,论文利用《方言调查字表》设计了一个包括文本语料和语音语料的西安方言语料库,录制了普通话和西安话平行的语音语料库。提出了基于归一化非线性多项式的方言韵律转换模型以及基于统计的方言时长转换模型和停顿时长转换模型。利用STRAIGHT算法修改普通话语音,实现普通话到西安话的转换。对转换结果的MOS评测表明,转换后的单字平均MOS得分4.60,双字平均MOS得分为4.75,语句的平均MOS得分为4.15。

英文摘要：

The conversion of dialect speech is an important research topic in the field of human-computer speech communication.A Xi＇an dialect corpus is built based on＂word-list in dialectal survey＂for Xi＇an dialect conversion from mandarin.Speech corpus is recorded with contrastive（Xi＇an dialect vs.mandarin） recordings.Prosodic models based on the normalized nonlinear polynomial method are built for Xi＇an dialect by analyzing the differences of pitch,duration and pause duration between Xi＇an dialect and mandarin.Xi＇an dialect is converted from mandarin by STRAIGHT algorithm.Subjective experiments demonstrate that the converted monosyllable,disyllable and sentence achieved 4.60,4.71 and 4.15 of the average Mean Opinion Score（MOS）.