论文部分内容阅读
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,从“维吾尔语语音声学参数库”中选择了包含边音/l/的单音节、双音节、三音节以及多音节词[0](即四音节以及四音节以上的词),提取它的声学参数并进行统计分析,探讨了边音的声学特性,分别研究了边音在单音节及多音节词中时的以及边音出现在词中位置(词首、词中、词末)时的共振峰、音强和时长分布模式,得出了一系列结论,其目的是为了提高语音合成的自然度即更好地为自然语言处理服务。本项研究结果维吾尔语语言乃至整个阿尔泰语系语言的韵律研究具有较高的参考价值。
According to the requirements of speech applications such as speech synthesis and recognition, starting with the textual analysis module, we selected the monosyllable, disyllable, trisyllable and polyphone containing polyphony / l / from the Uyghur Speech Acoustics Parameter Library Word [0] (that is, four syllables and four syllables above the word), extract its acoustic parameters and statistical analysis of the edge of the acoustic characteristics of the edge of the sound in the monosyllabic and polysyllabic words and A series of conclusions are drawn from the patterns of formant, pitch and duration when the edge occurs in the position of the word (prefix, word, end) and the purpose is to improve the naturalness of speech synthesis For natural language processing services. The results of this study Uyghur language and even Altaic language rhythm research has a higher reference value.