The emerging standards (such as HEVC) always improve coding efficiency by adopting more complex coding technologies which in turn greatly increase computational costs at the same time. However, in many application scenarios, available computational resources for encoder are constrained and varying, especially for mobile devices or real-time visual communication. Moreover, different devices are likely to have different computing capacities. Even for the same mobile device, the available computing capacities for video encoder vary from time to time because of multitasking. When computational resources are not enough to apply all the optional coding technologies, coding efficiency, i.e. R-D performance, will be reduced.

Different from conventional profiles in those emerging video coding standards and related optimization methods, we will focus on building the algorithms which are optimally scalable under constrained and varying computational capacity, in order to meet above requirements.
This talk will present a video coding framework based on priority order, and give the general thoughts of designing algorithms based on using the cost-performance of computational cost and coding efficiency gain as priority. We propose a feasible solution to this idea by introducing
a four level coding structure; as well as some novel features that can express the relationship between video contents and their priorities. The algorithm framework may be able to solve the realistic problems of optimally allocating computational capacity without affecting the ultimate compression efficiency.