Abstract: This paper presents a high performance VLSI architecture design for H.264/AVC entropy encoding. In the design, a fast encode all-zeros block method is proposed to accelerate the encoding process. A double RAM method is designed to avoid the mistake of storage and be easy to control data as a block. A pipeline technology is proposed to erase the RAM and make the process faster. The synthesis result shows the proposed architecture can improve the performance of the CAVLC.

Key words: Context based Adaptive Variable Length Coding(CAVLC), Exp-Golomb, H.264, Very Large Scale Integration(VLSI)

摘要： Exp-Golomb和CAVLC是H.264引入的新的熵编码形式，通过引入上下文的方式，减少编码码流，提高鲁棒性。该文提出一种熵编码的硬件结构，采用全0子块探测，双RAM结构，流水线技术，以及通过计算代替查找表的方法，加快编码过程，同时减少硬件的复杂度。FPGA综合结果显示，关键路径为11.449 ns，系统时钟最高支持到87.346 MHz。

关键词: 基于内容的变长编码, 指数哥伦布编码, H.264协议, 超大规模集成电路

CLC Number: 

• TN919.81