摘要
提出了一种DCT的新的无乘法硬件实现结构,用基于DA分布式算法的CORDIC旋转算法来实现DCT变换,它改变了单独使用DA分布式算法的寻址机制,减小了查找表的规模,使查找表读出的时间大大减少;而且能够避免单独使用CORDIC旋转算法需要的深流水线,它们一起实现速度和面积的共同优化。
This thesis presents a new architecture which has not multipliers, using a distributed arithmetic based on CORDIC algorithm for the computation of the 1D - DCT. It changes the mode of addressing ; this method decreases the scale of lookup table greatly and provides a small access time to the lookup tables. It can also avoid deep pipeline in order to decrease the size of chip, the two methods can provide a compromise between the area and the speed of the processor.
出处
《微处理机》
2008年第2期117-119,共3页
Microprocessors
