Implementation and Quality Assessment of a User- Centric Adaptation System for DASH

Abstract

In recent years, dynamic adaptive streaming over HTTP (DASH) has been widely used inorder to let the viewer have an uninterrupted streaming and therefore higher quality-ofexperience(QoE). The technique basically relies on dynamic rate adaptation by downloadingthe video at lower qualities according to various network conditions. However, manyissues are open related to client-side rate adaptation. For example, current DASH systemsdo not yet let the viewer decide which aspect of quality of the video should be reducedfirst and which aspect of quality of the video should be stable. This paper presents a noveladaptation system for DASH which is designed to search whether a higher QoE might beachieved if the adaptation method could be user-centric. During subjective quality assessment,it has been observed that all the participants have had a higher QoE comparedto regular DASH system when they were given a chance to decide the way of adaptation.