An approach to estimate correlation between QoS and perceptual video quality in packet-switched networks

Traditional QoS parameters estimated at network level are not suitable for describing quality of service perceived by the end user, but user experience can be described by QoE (Quality of Experience). QoS at user level is perceptual and subjective parameter, but there are recommendations which formalize procedures for testing perceptual quality. There are subjective and objective video quality assessment methods which are used for estimating video quality. This paper proposes a model for estimating the impact of QoS parameters on the perceptual quality of video signals. The proposed model assumes use of layered QoS model. QoS parameters of MPEG-4 video transmission in various network configurations were analyzed and by comparing estimated QoS parameters to perceived video quality a conclusion was made how some of the QoS parameters impact perceptual quality of video signals. EvalVid integrated into NS-2 simulator was used for estimating QoS parameters. SSIM index was used as objective video quality metric, which has been calculated using MSU VQMT tool.