ENHANCING VOICE QUALITY THROUGH IMPROVED E-MODEL: A MATHEMATICAL ANALYSIS

Abstract

The E-model was suggested by ITU-T G.107 standard, which is an ever changing test of calculations to evaluate the network performance whether it is ready to send a VoIP call or not. Several studies have been performed in this field and it highlights the complexity of this model; since a lot of lengthy mathematical calculation is needed because of several parameters which are available in network and it depends upon them. So a need of more simplified version occurs for estimating the voice quality. In this paper, we have assumed simple correction by using standard values to a simplified E-model and calculated the MOS value by taking the default correction coefficients values for 4 basic codecs which are G.711, G.723.1, G.726 and G.729A, and then we show that its predictions mapping with PESQ scores by implementing it with simple calculation. This paper focuses on enhanced VoIP voice call quality evaluation based on simplified E-Model so that the end to end conversation phase identified correctly and organization can be made. In the end, we performed various mathematical calculations to show the variance in MOS with respect to packet loss and better mapping compare to simplified model. Key Words: VoIP, E-model, PESQ, Voice quality, QOS.