A Novel Approach to Streaming QoE Score Calculation by Integrating Error Impacts

Video streaming services have become prominent in the last decade. As any other cloud service, these services are error-prone, and the errors during startup and/or playback affect the viewing experience of end-users. Hence, the calculation of Quality-of-Experience (QoE) scores should also account for error impacts. In this paper, we introduce a player-based error classification scheme, which classifies errors based on origin and severity. We use this scheme to quantify the quality degradation due to errors, and propose to improve the QoE score by integrating these quality factors. We instrument the open-source media players dash.js and Exoplayer in our proposed system which follows the guidelines of various multimedia streaming standards. We define several scenarios focusing on different QoE influencing factors, and assess our proposed model's performance. Comparisons with various state-of-the-art QoE models show that our model captures the effect on user experience better in scenarios induced with player-related errors.