论文部分内容阅读
本数据库是在分析、测量了藏语拉萨话733个单音节基础上形成的。全部声学数据分为声母(辅音)、核心元音、鼻音和声调4个数据群,共30项特征。为了使用上的方便,本数据库还配置了排序检索、统计分析和绘图功能。全部音节(并数据)可按声母、核心元音、韵尾和声调不同的优先等级排出次序,也可任意检出某个音素或声调。统计分析包括了平均值、标准差、标准误、置信区间估计等算法。实践证明,语音实验研究中利用数据库,不仅有利于对特定语言的语音系统进行定量描写,而且在语音合成中也能提供必要的参考数据。
This database was formed on the basis of analyzing and measuring 733 monosyllabic syllables in Tibetan Lhasa dialect. All acoustic data are divided into consonants (consonants), core vowels, nasal and tone data of four groups, a total of 30 features. In order to use the convenience, the database is also configured to sort the search, statistical analysis and drawing capabilities. All syllables (and data) can be out of order according to the different priority of the initials, core vowel, ending and tone, but also a phoneme or tone can be detected at will. Statistical analysis includes the average, standard deviation, standard error, confidence interval estimation algorithm. Practice has proved that the use of the database in the voice experiment research is not only conducive to quantitative description of the voice system of a specific language, but also can provide the necessary reference data in the voice synthesis.