DFT spectrum filtering

Author

Zarowski, Christopher J. ; Yunik, Maurice ; Martens, G.O.

Author_Institution

Dept. of Electr. Eng., Manitoba Univ., Winnipeg, Man., Canada

Volume

36

Issue

4

fYear

1988

fDate

4/1/1988 12:00:00 AM

Firstpage

461

Lastpage

470

Abstract

A.E. Kahveci and E.L. Hall (see IEEE Trans. Comput., vol.C-23, no.9, p.976-81, Sept. 1974) introduced the concept of filtering discrete Fourier transform (DFT) spectra in the Walsh sequency domain. This is accomplished by finding a real and block-diagonal Walsh filter matrix G_w in the Walsh domain that performs the sample filtering operation as the prototype complex diagonal Fourier filter matrix G_f in the Fourier domain. The present authors provide additional information on the structure of G_w and generalize some results by C.J. Zarowski and M. Yunik (see ibid., vol.ASSP-33, p.1246-52, Oct. 1985). They consider a more general class of transforms, the T transforms, and the structure of the resulting T transform filter matrices G _t. Examples of T besides T=W are considered, such as the Harr transform and fourth-order Chrestenson transform. The implementation of the presented DFT spectrum filtering techniques using linear systolic arrays is briefly considered

Keywords

fast Fourier transforms; filtering and prediction theory; spectral analysis; DFT spectrum filtering; Fourier; Fourier domain; Harr transform; T transform filter matrices; T transforms; Walsh domain; Walsh sequency domain; block-diagonal Walsh filter matrix; complex diagonal Fourier filter matrix; fourth-order Chrestenson transform; linear systolic arrays; real Walsh filter matrix; Convolution; Councils; Discrete Fourier transforms; Discrete transforms; Filtering; Nonlinear filters; Prototypes; Systolic arrays; Vectors; Very large scale integration;

fLanguage

English

Journal_Title

Acoustics, Speech and Signal Processing, IEEE Transactions on

Publisher

ieee

ISSN

0096-3518

Type

jour

DOI

10.1109/29.1550

Filename

1550