How to Low-Pass filter using Millipede´s FFT?

Dear experts,

I´m trying to low-pass filter a 1-Dimensional random signal using the 1D Fast Fourier Transform (FFT) component in Millipede. However, despite the FFT component seems to work, the Inverse FFT can´t reconstruct the original signal as expected. Please, note that the “Original Signal” and the “Inverse FT” graphs are not very similar either in the figure or in this simple test file: FFTfilter.gh (32.5 KB).

1. Am I doing something wrong?
2. Any alternative to Low-Pass filter a signal in GH?

PS: A low-pass filter is a natural way to control spatial resolution of random noise for some later application, e.g. to randomly perturb some structure for non-linear buckling analysis in Karamba.

Does anybody have some ideas about Fourier Transforms in GH? Thanks!!

you could use a Graphmapper (for filtering/bandwidth filtering) - this is me trying to bridge my experience in FLStudio into GH - one of many doors opened today.

A synch wave might work, but a oh, wow*, try any wave you like! You can chain graphmappers to go further - and if you search for - ok, now i remember where i saw this:

Must be tailored to your need but you can change the bandwidth, filter high/low (of the sine wave, your goal)…

Is live audio I/O (recording to file/ram) possible in GH?

1 Like

ok, here you have an excentric High/Low…

1 Like

ok, sorry i went off the subject, reviewing FFTs again.

Hi…two m-functions, sfft and sifft, are supplied as plain text counterparts to the Matlab binary built-in functions of fft and ifft. You can not only read their codes but also can step through their executions. The accompanied document derives and explains the Cooley-Tukey fast Fourier transform (FFT) algorithm.

1 Like