FPGA implementation of fully parallel fast MDCT algorithm

Abstract

Compression of digital audio signals has become increasingly more important with the advent of fast and inexpensive microprocessors and digital signal processors. Several compression/decompression schemes were developed and well estabilished. Most of them adopt the modified discrete cosine transform (MDCT) and its inverse version (IMDCT). In this paper we present a FPGA implementation of a fast rotation-based forward and inverse MDCT algorithm. We implemented fully parallel IP core capable of both forward and inverse 36-point MDCT computation. It was found that the computation is faster than in well known FPGA implementations. The core is suitable for the very high-performance demanding multi-channel audio applications.