摘要
蒙古语电话语音的研究刚刚起步,建立规模较大的、通用的电话语音语料库成为了重要的基础工作.本次建立的蒙古语电话语音语料库是以自然口语的对话形式进行录制,并且体现了不同电话信道、不同方言、不同年龄段说话人的特点.本文详细讨论了语料库的录制整理、语音切分和语音标注等几个问题.本次建立的语料库为蒙古语电话语音的语音识别、语音检索、语音监控和说话人识别等技术的研究提供了真实的实验数据.
Study of Mongolian telephone speech is at the early-stage and constructing a larger, u- niversal telephone speech corpus is an important fundamental work. A Mongolian telephone speech corpus is constructed, which recorded in the form of natural spoken dialogue and reflected the fea- tures of different telephone channels, dialects and ages. Several issues including recording and colla- ting of the corpus, speech segmentation and speech tag are discussed. This corpus will provide Mon- golian telephone speech data in a real environment for the researches of speech recognition, speech retrieval, speech monitoring and speaker recognition.
出处
《内蒙古大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第3期320-323,共4页
Journal of Inner Mongolia University：Natural Science Edition
基金
内蒙古自然科学基金重大项目(2011ZD11)
国家自然科学基金项目(No.61263037
No.71163029)
国家电子信息产业发展基金项目资助
