On Sun, 16 Feb 2014 02:45:52 -0800 (PST), martin.mullins942@gmail.com wrote: [Snipped by Lyons] Sorry for not replying to you sooner.>I'm zero padding from a 2^n to 2^n, say K=2048 window length and N = 8192 the fft length. > >After the suggestions from this article it works but only for the case >of zero padding only after the original data rectangle window. >ie. x(n) = [data (len: K)][zeros (len: N-K)] -- (1) >If X(m) = FFT{x(n)} >Then X_window(m) = alpha*X(m) - beta/2 * X(m-N/K) - beta/2 * X(m+N/K) >So the only difference is that instead of X(m +/- 1) terms its X(m +/- N/K) >(Was determined by taking the FFT of the hanning window).You wrote: "X_window(m) = alpha*X(m) - beta/2 * X(m-N/K) - beta/2 * X(m+N/K)" I don't think that's correct, but I'll have to do some software modeling to be sure. Your original question is an interesting question, and I intend to find the correct answer to your question. [-Rick-]

# How do apply window in the Frequency Domain for a signal that was zero padded before FFT?

Started by ●February 10, 2014

Reply by ●March 1, 20142014-03-01