AN EFFICIENT DESIGN OF CONTEXT MODELER FOR CABAC ENCODER IN H.264

Context-based adaptive binary arithmetic coding (CABAC) is a very important entropy coding technique defined in H.264. It is not easy to store and retrieve efficiently the previously encoded syntax elements especially when full hardwired VLSI implementation of CABAC is required. This paper presents an efficient design of context modeler which uses the well-organized location-index assignment and the copy-and-paste technique to locate, store and retrieve the needed syntax elements both easily and efficiently with less storage space. Synthesis results show that our CABAC encoder occupies 19339 gate counts and operates at 200 MHz by using the TSMC 0.13 mu m CMOS technology. In the simulation, our design can perform the throughput of 166 Mbps, which is able to support H.264 CIF encoding in real time.