Title :
DCT algorithms for composite sequence lengths
Author :
Bi, Guoan ; Yu, Lee W.
Author_Institution :
Sch. of Electr. & Electron. Eng., Nanyang Technol. Inst., Singapore
fDate :
3/1/1998 12:00:00 AM
Abstract :
This paper presents fast algorithms for the type-II and -III discrete cosine transforms of composite sequence length. In particular, a radix-q algorithm, where q is an odd integer, is derived for uniform or mixed radix decomposition of the discrete cosine transform. By combining the radix-q and radix-2 algorithms, a general decomposition method for any composite length is developed. Reduction of computational complexity can be achieved for many sequence lengths compared with that needed by the well-known radix-2 algorithm. Furthermore, both the proposed and Chan and Siu´s (1993) mixed radix algorithms achieve the same computational complexity for N=3*2p and 9*2P. However, our algorithm uses a simpler decomposition approach and provides a wider range of choices of sequence lengths
Keywords :
computational complexity; digital arithmetic; discrete cosine transforms; sequences; signal processing; DCT algorithms; composite sequence lengths; computational complexity; mixed radix decomposition; radix-2 algorithm; radix-q algorithm; type-II discrete cosine transforms; type-III discrete cosine transform; uniform radix decomposition; Bismuth; Computational complexity; Computational efficiency; Costs; Digital signal processing; Discrete cosine transforms; Equations; Helium; Partitioning algorithms; Signal processing algorithms;
Journal_Title :
Signal Processing, IEEE Transactions on