Analysis of H.264/AVC Scalable Video Coding for Video Delivery to Heterogeneous Terminals

In 2007, the ITU-T H.264 | ISO/IEC MPEG-4 AVC standard was extended to support temporal, spatial and fidelity (SNR) scalability in a framework that is referred to as Scalable Video Coding (SVC). Since the development of this SVC extension, its use has been proposed for several applications. It seems however there is not yet a broadly agreed understanding about the benefits of SVC compared to non-scalable coding. In this paper, we describe coding efficiency gain and cost measures for scalable video against non-scalable simulcast, and single layer non-scalable coding, respectively, in the context of video delivery to heterogeneous terminals. Our results show that the cost and gain from SVC are strongly dependent on the application and conditions. Specifically, it is shown that while SVC can theoretically provide promising gain in some applications, its cost is not negligible and in some cases this cost can outweigh the gain.