摘要
本文讨论了基于MFCC(Mel-Frequency Cepstral Coefficients)听觉感知模型的语音质量客观评估方法(EPESQ:Efficient Psychoacoustics Evaluation of Speech Quality)在网络语音质量客观评估中的应用,深入分析了EPESQ算法的基本原理,并对算法的MFCC语音特征参数进行了改进和性能分析。实验证明该算法在确保评测准确性的前提下,大大降低了算法的复杂度,具有较强的适应性,可靠性,实用性,完全可以应用于通信网络语音质量评估。
The paper discusses EPESQ algorithm based on MFCC which is used in quality evaluation of communication networks. The basic principle is studied extensively, MFCC is improved, and the performance of the algorithm is analyzed. Experimental results show that the new algorithm retains the comparable performance and reduces the complexity. It has a strong adaptability, reliability, practicality and can be applied in quality evaluation of communication networks.
出处
《电路与系统学报》
CSCD
北大核心
2009年第3期111-116,共6页
Journal of Circuits and Systems