Personalizing Quality Aspects in Scalable Video Coding

In video coding, certain limitations imposed by the environment, most typically the bit rate, need to be fulfilled. This is achieved by allowing the encoder to reduce the quality in one or several ways, such as the distortion, the resolution and the frame rate. The upcoming scalable video coding mechanisms allow this reduction to take place not exclusively during the encoding step, but at any time. This allows us to reduce the quality in a more personalized way, taking into consideration the preferences of the end user. This paper presents a framework that enables such user dependent quality reductions. We validated this framework by means of a test involving 19 test persons. The results of this mechanism are good, but up to now not sufficiently reliable to use it in commercial applications. At the same time, we still see some room for improvement