RASTA滤波在语音干扰客观评价中的应用

Application of RASTA Filter in Speech Communication Interference Result Evaluation

提出了相关频谱(Relative Spectral,RASTA)处理模拟人类听觉感知特性的原理,将RASTA处理应用于语音通信干扰效果的客观评价。在美尔倒谱系数(Mel Frequency Cepstral Coefficient,MFCC)参数的基础上,加入RASTA处理步骤,对语音通信干扰试验中的实测数据进行分析,证明了RASTA处理在一定程度上改善了MFCC参数的性能。

This paper presents the principles of imitating mankind sense of hearing features by using RASTA （Relative Spectral） processing and applies RASTA filters to the objective evaluation of speech communication interference effect. On the foundation of MFCC （Mel Frequency Cepstral Coefficient） parameter,it adds RASTA processing steps and analyzes the sound communication interference data acquired from the outfield experiment. The results show that RASTA filters can improve the features of MFCC parameter to some extent.