H.264/AVC自适应算术解码器结构设计

An Architecture Design for Context-Based Adaptive Binary Arithmetic Decoder in H.264/AVC

限制基于上下文的二进制算术解码(CABAD)速度的几个主要环节入手,提出了优化的上下文存储模式,改进的重归一化单元,并使用流水线提高解码速度.在Synopsys公司的CoCentric System Studio平台进行了二进制算术解码器体系结构设计,仿真结果表明,本结构能够满足主要档次(main profile)CIF 30fps的实时解码的要求.

In this paper, on the base of dealing with the steps that limit the speed of CABAD, an optimized context model maintenance and an improved renormalization are proposed, and with the use of pipline the decoding speed is improved. The proposed architecture was designed and validated on the CoCentricSystemStudio platform provided by Synopsys Co., and results show that the need for real time decoding of main profile CIF 30fps bitstrearn is satisfied.