论文部分内容阅读
目前在蒙古语标准音水平测试中主要存在两个问题:一是工作量大,效率低;二是测试手段尚未完善,很难确保测试结果的准确性和公正性。为了提高工作效率和测评质量,我们提出了建立“蒙古语标准音语音数据库”开发和研制机器测评系统的思路。本文阐述了建设“蒙古语标准音语音库”的初步构思,如文本库、原始语音库和语音数据库等的设计和建立以及开发蒙古语标准音机器测试系统。“蒙古语标准音语音数据库”不仅对蒙古语标准音机器测试系统的研究和开发提供大量的语音素材和科学的声学参数数据以及相关语音知识,而且对蒙古语语音合‘成与语音识别研究及人机对话等言语工程领域具有重要理论意义和现实应用价值。
At present, there are mainly two problems in Mongolian standard sound level test: one is the heavy workload and the low efficiency; the other is that the testing method is not yet perfect and it is difficult to ensure the accuracy and fairness of the test result. In order to improve work efficiency and quality evaluation, we put forward the idea of establishing a “standard Mongolian standard voice and speech database” to develop and develop a machine evaluation system. This article elaborates on the initial idea of building a “standard Mongolian standard audio speech library”, such as the design and construction of a text database, original speech database and speech database, and the development of a Mongolian standard audio machine testing system. “Mongolian standard pronunciation voice database ” not only provides a large amount of voice material and scientific acoustic parameter data and related phonetic knowledge for the research and development of Mongolian standard sound machine testing system, Research and man-machine dialogue and other speech engineering has important theoretical significance and practical value.