Abstract
This paper describes efficient hardware architecture for the deblocking filter used in H.264/AVC baseline profile video coding standard. The deblocking filter is a computationally and data intensive tool leading to an increased execution time of both encoding and decoding processes. In fact, we propose a novel edge filter ordering which needs 64 clock cycles to filter a Macroblock (MB). A specified memory organization is also applied in order to avoid unnecessarily waiting for availability of the pixels that will be filtered. The proposed architecture includes both pipelining and parallel processing techniques and is implemented in synthesizable HDL. This hardware is designed to be used as module of a complete H.264/AVC decoder which the functionality was validated on Nios II at 100 MHz.