Adaptive Video Streaming through Estimation of Subjective Video Quality

Concerning video transmission on the Internet, we present a model for estimating the subjective quality from objective measurements at the transmission receivers and on the network. The model reflects the quality degradation subject to parameters like packet loss ratio and bit rate, and is calibrated using the prerecorded results from subjective quality assessments. Besides the model and the calibration, the main achievement of this paper is the model’s validation by implementation in a monitoring tool. It can be used by content and network providers to swiftly localise the causes of a poor quality of experience (QoE). It also can help content providers make decisions regarding the adjustment of vital parameters, such as encoding bit rate and error correction mechanisms. We show how the estimated subjective service quality can be applied for decision making in content delivery networks that consist of overlay networks and multi-access networks. Keywords—Quality of Experience; perceived quality; video streaming; assessment; adaptation