Newly developed HTTP-based video streaming technologies enable flexiblerate-adaptation under varying channel conditions. Accurately predicting theusers' Quality of Experience (QoE) for rate-adaptive HTTP video streams is thuscritical to achieve efficiency. An important aspect of understanding andmodeling QoE is predicting the up-to-the-moment subjective quality of a videoas it is played, which is difficult due to hysteresis effects andnonlinearities in human behavioral responses. This paper presents aHammerstein-Wiener model for predicting the time-varying subjective quality(TVSQ) of rate-adaptive videos. To collect data for model parameterization andvalidation, a database of longer-duration videos with time-varying distortionswas built and the TVSQs of the videos were measured in a large-scale subjectivestudy. The proposed method is able to reliably predict the TVSQ of rateadaptive videos. Since the Hammerstein-Wiener model has a very simplestructure, the proposed method is suitable for on-line TVSQ prediction in HTTPbased streaming.
展开▼