One spoonful or multiple drops investigation of stalling distribution and temporal information for quality of experience over time

The interruptions of playback fluency are considered to be essential phenomena in the degradation of video quality at the user's side during the consumption of HTTP streaming services. Indeed, rebuffering can certainly negatively affect the overall experience, however, the extent of annoyance depends on the content and also on the parameters of stalling, like frequency or duration. In this paper we present a subjective video quality assessment measurement in which a series of videos containing different distributions of frame freezing were evaluated by human observers. The selected points of rebuffering were chosen based on their Temporal Information, which can reflect motion content.