A quest for an Internet video quality-of-experience metric

摘要: An imminent challenge that content providers, CDNs, third-party analytics and optimization services, video player designers in the Internet ecosystem face is lack of a single "gold standard" to evaluate different competing solutions. Existing techniques describe quality encoded signal or controlled studies measure opinion scores do not translate directly into user experience at scale. Recent work shows measurable performance metrics such as buffering, startup time, bitrate, number bitrate switches impact experience. However, converting these observations quantitative quality-of-experience metric turns out be challenging since are interrelated complex sometimes counter-intuitive ways, their relationship can unpredictable. To further complicate things, many confounding factors introduced by nature itself (e.g., interest, genre). We believe issue interdependency addressed casting this machine learning problem build suitable predictive model from empirical observations. also show setting up based on domain-specific measurement-driven insights minimize various improve prediction performance.