节点文献
基于云理论的支持无输入的语音质量客观评价的建模方法
The Modeling Method for Non-intrusive Objective Speech Quality Measurement Based on Cloud Theory
【Author】 Li Jing,Li Lingling,Wan Juan,Xu Genhua,Chen Fang School of Computer Science,Wuhan University,Wuhan 430079,P.R.China
【机构】 武汉大学计算机学院;
【摘要】 语音质量客观评价体系是语音处理中一个重要研究内容,而基于输入——输出的语音质量客观评价在实际中存在很多不足。本文提出一种基于云理论的支持无输入语音质量客观评价的建模方法。在该模型中给出了规则发生器的构造,并提取特征参量和进行基于主观质量数据库的语音质量测试,并进行客观系统与主观评价结果的相关度分析。实验结果表明该方法能够有效提高语音质量评价的准确度和降低语音质量评价的复杂度。
【Abstract】 Objective speech quality assessment is a important research in speech processing recently,but there are so many disadvantages in assessment based on input-output.A novel modeling method for non-intrusive objective speech quality measurement based on cloud theory is proposed in this paper.The mothod for rule-maker constructed and feature ex-tracted is proposed.The speech quality for subjective quality speech database is measured.The correlation between subjective system and objective system is analyzed.The Experimental results show that the method is an effective technique to increase nicety and decrease complexity for speech quality assessment.
【Key words】 Non-intrusive; Objective Speech Quality Assessment; Cloud Theory;
- 【会议录名称】 第二十六届中国控制会议论文集
- 【会议名称】第二十六届中国控制会议
- 【会议时间】2007-07-26
- 【会议地点】中国湖南张家界
- 【分类号】TN912.3
- 【主办单位】中国自动化学会控制理论专业委员会(Technical Committee on Control Theory,Chinese Association of Automation)