Effect of network QoS on user QoE for a mobile video streaming service using H.265/VP9 codec

Online video streaming is one of the most popular yet challenging services being used today where the Internet Service Providers (ISP's) need to guarantee certain assured service quality level for the end user. However, as the data transmission takes place over an unreliable error-prone channel; it is a challenging task for the ISP's. Normally, the user perceived quality is best expressed in terms of the Quality of Experience (QoE) factor or Mean Opinion Score (MOS) factor. On the contrary, the various factors that affect the video streaming quality; commonly referred to as the Quality of Service (QoS) factors, are relatively easier and economic to be measured by the service providers. In this paper, we consider five network layer specific QoS impairment factors (variable initial delay, packet loss, jitter, bandwidth and buffering delay) and try to assess their impact on the QoE by considering each one of them separately. We use a non-linear regression approach to compare the various QoS to QoE mapping models and find out the most suitable/optimal model for our individual cases based upon a decision variable.