Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
156 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

A class of AM-QFT algorithms for power-of-two FFT (1404.1810v1)

Published 7 Apr 2014 in cs.DS and cs.NA

Abstract: This paper proposes a class of power-of-two FFT (Fast Fourier Transform) algorithms, called AM-QFT algorithms, that contains the improved QFT (Quick Fourier Transform), an algorithm recently published, as a special case. The main idea is to apply the Amplitude Modulation Double Sideband - Suppressed Carrier (AM DSB-SC) to convert odd-indices signals into even-indices signals, and to insert this elaboration into the improved QFT algorithm, substituting the multiplication by secant function. The 8 variants of this class are obtained by re-elaboration of the AM DSB-SC idea, and by means of duality. As a result the 8 variants have both the same computational cost and the same memory requirements than improved QFT. Differently, comparing this class of 8 variants of AM-QFT algorithm with the split-radix 3add/3mul (one of the most performing FFT approach appeared in the literature), we obtain the same number of additions and multiplications, but employing half of the trigonometric constants. This makes the proposed FFT algorithms interesting and useful for fixed-point implementations. Some of these variants show advantages versus the improved QFT. In fact one of this variant slightly enhances the numerical accuracy of improved QFT, while other four variants use trigonometric constants that are faster to compute in `on the fly' implementations.

Summary

We haven't generated a summary for this paper yet.