查看文章

The paper deals with a non-intrusive approach to the speech quality assessment and examines an impact of network jitter on effective equipment impairment factor in E-model. In order to improve currently used computational E-model (ITU-T G. 107), we proposed an optimization based on numerous measurements and experiments with an objective intrusive method PESQ (Perceptual Evaluation of Speech Quality) under varying delay, packet loss, jitter and play-out buffer. Comparing achieved results in experiments and the computed values in E-model, we proposed modification which improves estimated MOS of E-model.