A quality of experience model for Adaptive Media Playout

In the past decade Adaptive Media Playout (AMP) has been intensively studied with respect to the detection of when to increase or decrease the playback rate in order to maintain a certain buffer fill state. In this paper we subjectively assess the QoE of AMP with respect to non-periodically and randomly selected content sections of a video sequence by using crowdsourcing. Furthermore, we introduce metrics that allow to quantify the distortion for audio and video that are caused by increasing or decreasing the playback rate. With these preliminaries we study the correlation between the introduced metrics and the subjectively assessed QoE. Therefore, we derive a utility model that allows estimating the QoE with the introduced metrics. We instantiate and validate the model using the data gathered from the conducted study.