A Method for Detecting Noise, Single Pass, Cross Pass and Echo of Mobile Phone

Quality of Experience is referred to as an important index for measuring how well the user's requirements are satisfied, and many methods have been presented to evaluate speech quality. But they cannot identify the kind of disturbances when the quality of speech is not good. This paper aims at proposing a method which can identify the kind of disturbances lowering the Quality of Experience. The approach proposed is based on PESQ and P563 algorithms. It adopted asymmetrical factor of PESQ and cosine distance in frequency domain to compare original signal with degraded signal. And a new time realignment method and temporal masking method were used to improve detection precision in wireless environments. This new algorithm is tested in real world including GSM and UMTS environments. The results indicate that this algorithm is helpful for detecting noise, single pass, cross pass and echoing automatically.