Skip to Main Content
Associating DVB-T2 and Scalable Video Coding (SVC) constitutes an efficient way for broadcasting added-value video services, such as HDTV and 3D TV, to end-users. The ultimate objective of this new approach of broadcasting video services is to ensure high Quality of Experience (QoE) for end users. Whilst Quality of Service (QoS) is the collective effect of performance that determines the degree of satisfaction of a user of a service, QoE reflects more accurately the user experience, as it is based on human perception when evaluating the video quality. Maximizing user QoE is thus becoming a crucial requirement when deploying new broadcast platforms for the provisioning of high quality video services. The contributions of this paper are two-fold. At first, we introduce a reference-less QoE measurement tool dedicated to SVC coding. Based on a learning function, this tool is able to learn the non-linear relationship between parameters affecting video quality and perceived user QoE. According to several experiments carried out using this tool, we demonstrate that decoding all SVC layers is not always efficient to achieve high user QoE, mostly when SVC enhanced layers experience packet losses. For the sake of maintaining a good QoE, it is worthwhile withdrawing enhanced layers experiencing packet losses and not displaying them to end-users. Based on this observation, we propose a QoE-Based Adaptive SVC Decoding (QoE-BASD) algorithm that assists a video receiver to select the appropriate SVC layers for video decoding in order to maximize QoE. We evaluate the performance of the proposed solution: (i) analytically, by using discrete Markov Chains to model the proposed solution; and (ii) via OPNET-based computer simulations. The obtained results are encouraging, and illustrate the gain achieved by QoE-BASD when compared to the conventional approach.