Talk:Rader's FFT algorithm

Hey, does anyone have a location for where this conference took place. The citation reads:

C. M. Rader, "Discrete Fourier transforms when the number of data samples is prime," Proc. IEEE 56, 1107–1108 (1968)

but I don't see a location.


 * Proceedings of the IEEE is a journal, not a conference. See here. —Steven G. Johnson 21:54, 2 April 2007 (UTC)

Finding the generator (g)
The algorithm explained in the article uses a generator - g - of the modulo N multiplication group, known to exist from number theory. However, no algorithmic way is mentioned to find such a generator. Is there an efficient way to do this? 2A02:8109:9340:112C:FD62:EAA0:5CCE:62F8 (talk) 01:01, 9 February 2015 (UTC)


 * Since the generators are extremely common, just exhaustive testing will turn one up pretty quickly, although there are slightly faster algorithms than this. (See e.g. Donald E. Knuth, The Art of Computer Programming, vol. 2: Seminumerical Algorithms, 3rd edition, section 4.5.4, p. 391 (Addison–Wesley, 1998).)  I'll add a link.  (It's also discussed in Primitive root modulo n.) — Steven G. Johnson (talk) 18:03, 24 October 2017 (UTC)

So who is "Rader"?
This article isn't very clear who the "Rader" in question is. If it's Rader's FFT algorithm, then it should say who it was named after.--Varkman (talk) 05:16, 20 November 2015 (UTC)


 * Note that this has been fixed (it is Charles M. Rader, a well-known figure in the signal-processing community). — Steven G. Johnson (talk) 17:52, 24 October 2017 (UTC)

Time complexity without zero-padding
Without zero-padding, the article states that the worst case requires O(N2) time. But each step of the recursion halves N, therefore isn't it still O(N log N) time? — Preceding unsigned comment added by Raysphere24 (talk • contribs) 08:18, 3 January 2021 (UTC)


 * Agree. 128.54.13.249 (talk) 22:14, 9 November 2023 (UTC)
 * Agree as well. I see no O(n2) runtime. 2A02:1210:2657:D600:63B7:3ED2:BBA4:60D3 (talk) 10:35, 23 May 2024 (UTC)