| The Discrete Fourier Transform (DFT) and Discrete Cosine Transform (DCT) have been widely used in many digital signal processing areas such as image processing, speech processing, spectrum analysis, etc. In particular, the 2-D DCT has been adopted in many international image compression standards such as MPEG, JPEG and CCITT. Although there has been great interest in fast algorithms and hardware architectures for the DFT and DCT, only few approaches meet the technological constraints for the design of architecture suitable for VLSI implementation.; The purpose of this research is to develop new architectures and implementation methods for the multidimensional DFT and DCT appropriate for VLSI implementation. The new implementation methods are based on various computer arithmetic techniques such as CORDIC algorithm and distributed arithmetic. These techniques are exploited to reduce area, increase throughput and get simple and regular designs. Simulation and performance evaluation of the proposed signal processors are conducted along with a fixed-point error analysis. |
