RtP/RTCP Streaming Video Qoe Assessment Under Network Impairments

As one of the most widespread used applications, streaming video services’user subjective quality perception has gained plenty of attentions by service providers.User subjective quality perception is mainly evaluated in two ways. One is subjective quality score, which use the Mean Opinion Score (MOS), a subjective quality measure acquired by averaging scores (on the scale of1to5) from a number of observers (assessors), to characterize Quality of Experience (QoE) of streaming videos. The other one is objective quality score, depending on subjective quality score data and scientific modeling process to assessment QoE of streaming videos based on video objective parameters and network service parameters.In this paper, a no-reference method, considering degraded video content and network QoS parameters, is proposed. Our proposed objective quality assessment model is derived through recording users’subjective perception on streaming video samples in subjective quality assessment phase and then summarizing regular pattern of human perception by nonlinear regression. Before deriving the QoE model formula, spatial and temporal characteristics, which rely on video content, affect video robustness of packet loss and help to improve the adaptability to different videos, are extracted separately. In quality assessment modeling phase, encoding compression impairment on video quality is evaluated relying on video bit rate and spatial-temporal characteristic, and packet loss quality degradation on streaming videos is assessed depending on packet loss and spatial-temporal characteristic. To verify the accuracy of our QoE model, validation samples are introduced to predict subjective perception.The Pearson Correlation Coefficient (PCC) of all samples is as high as0.966and the Root Mean Square Error (RMSE) is0.289. Therefore, our proposed model has an excellent performance of predicting subjective perception of streaming videos and can provide good algorithm support for network service provider to detect streaming video’s QoE score.Currently, this streaming video objective quality evaluation model has applied to the detection part of China Telecom’s video center.