A VLSI Architecture for Reference Frame Compression on High Definition Video Coding Systems

This paper presents a hardware-friendly solution for external memory bandwidth reduction focusing in the state-ofart video encoders, like H.264/AVC and HEVC. The proposed approach is based on reference frame compression using an adaptation of the traditional Huffman algorithm and is a lossless solution. This solution is entitled RFCAVLC8t (Reference Frame Context Adaptive Variable-Length Coder with 8 tables) and is based on the use of static Huffman tables. The RFCAVLC8t is a hardware-friendly version of the Huffman algorithm and it uses eight static tables to avoid the cost of the on-the-fly Huffman statistical analysis. The best table to encode a block is defined using a context evaluation, resulting in a context-adaptive configuration. The use of RFCAVLC8t reached an average compression rate higher than 31% for the evaluated video sequences. The RFCAVLC8t hardware (Coder and Decoder) was designed and synthesized to an FPGA device. The RFCAVLC8t design is able to reach real-time encoding for WQSXGA (3200 x 2048 pixels) at 30 fps.