On the realization of discrete cosine transform using the distributed arithmetic

Abstract

In this paper, we propose a unified approach for the realization of forward and inverse discrete cosine transforms. By making use of this approach, one can realize an odd prime length DCT/IDCT with two half-length convolutions without extra overheads in terms of the number of multiplications. This formulation is most suitable for the realization using the distributed arithmetic. In such a case, typical convolvers can be used as the core unit for the hardware implementation of the transforms. Hence, an efficient unified DCT/IDCT chip can be designed. A 2-D 11 x 11 unified DCT/IDCT chip is also proposed to demonstrate the superiority of the proposed formulation in this paper. The proposed architecture can easily meet the speed requirement of 14.3-MHz real-time operation with the current 2-μm CMOS technology.