Derivation of Diagonalized Identical Equations for 3 by 3 Circulant and Quasi-Circulant Matrices and Their Application to Winograd 7-Point FFT

The derivation of the 7-point Winograd fast Fourier transform (FFT) requires complex steps such as using the Rader prime algorithm to turn an N-point discrete Fourier transform (DFT) into an (N − 1)-point convolution and then using the Chinese remainder theorem for polynomials to find the set of rem...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of Signal Processing 2022/01/01, Vol.26(1), pp.13-19
Hauptverfasser: Takahashi, Nobuaki, Takago, Daisuke
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The derivation of the 7-point Winograd fast Fourier transform (FFT) requires complex steps such as using the Rader prime algorithm to turn an N-point discrete Fourier transform (DFT) into an (N − 1)-point convolution and then using the Chinese remainder theorem for polynomials to find the set of remainders. In this paper, we describe a simpler method for deriving the 7-point Winograd FFT using the diagonalized identical equations of 3 by 3 circulant and quasi-circulant matrices. These diagonalized identical equations of 3 by 3 matrices are not found in the literature and are newly derived.
ISSN:1342-6230
1880-1013
DOI:10.2299/jsp.26.13