The Research and Development of a Real-Time Qoe Evaluation System for VI Deo Streaming Services

For video streaming services, at the client side the playout interruption, video blocks and unpredictive preroll time may occur during a video playout due to the random delay and jitter encountered in the Internet. In addition, the video playout rate variation under an Adaptive Media Playout (AMP) mechanism may also affect the user perception. In general, one can obtain the Quality of Service (QoS) parameters, e.g., packet losses, playout delay and underflow time, by using network measurement equipments or programs. However, these QoS parameters can not reveal the users’perception of the video streaming services. Consequently, the subjective test need be employed in order to understand the users’perception. However, the subjective test is time-consuming and rather costly. Thus, a cost-effective and real-time QoE evaluation method for video streaming services is presented in this proposal. The proposed scheme first adopts subjective tests to find the relationships between the QoE and individual QoS parameters. Next, based on subjective test results, the corresponding QoS-QoE mapping functions for individual QoS metrics are derived by using the regression approach. Finally, the integrated QoE function, which is a multi-variate function of multiple QoS metrics, for assessing the overall QoE of video streaming services is proposed. Hence, network/service providers only need to measure related objective QoS parameters while the corresponding MOS/QoE can be easily derived in real-time based on the integrated QoE function. Since subjective tests only need to be conducted occasionally for updating the parameters in the QoS-QoE mapping functions, the cost of QoE assessment can decrease significantly. Finally, we will develop a real-time QoE measurement system by software.

Project ID:PB10108-2801
External Project ID:NSC101-2221-E182-004