A few comments: If you take the FFT of an FFT of a time signal, the result is a time-reversed version of the original. That's the main difference between an FFT and an inverse FFT: the inverse FFT fixes this by reversing time.
The combination of FFT and inverse FFT raises amplitude by a factor of N, if amplitude is uncorrected. The correction is often built into the inverse FFT so it's fixed transparently, by dividing every sample by N. However, you also often see no correction in the inverse FFT, so your amplitude out is raised by a factor of N. Occasionally you see correction of a factor of 1/sqrt(N) applied to both forward FFT and Inverse FFT, which makes them more symmetric and also fixes the amplitude gain. For many PFBs, you *can* simulate the effects of a filter by multiplying the PFB outputs by the frequency response of the filter. The results aren't absolutely perfect, but they can be very good, and they approach perfection as the number of PFB subbands approaches infinity. Regards, Ross On Thu, Apr 23, 2020, 3:05 AM James Smith <jsm...@ska.ac.za> wrote: > Hello Heystek, > > This should have been part of your undergrad signal-processing course. > Usually in an inverse FFT, they include a scaling factor of 1/sqrt(2) or > something like that, I forget exactly. You might find some references here: > https://www.dspguide.com/ > > The application is doing signal-processing in the frequency domain, which > can often be computationally more efficient than in the time-domain. You'd > multiply a window of frequency domain data with the frequency-domain > representation of what your filter should look like, then IFFT to get a > filtered time-domain signal. > > Just note that for a spectrometer we'd generally use a PFB, not just an > FFT, so the same mathematical relation is NOT true for these cases. > > Regards, > James > > > On Thu, Apr 23, 2020 at 10:00 AM Heystek Grobler <heystekgrob...@gmail.com> > wrote: > >> Hey James >> >> I thought of the “to workspace” sink. I am not to familiar to write from >> Simulink to the workspace, but I will give it a go! Thanks for the help. >> >> Out of curiosity, if I have both halves of the symmetric FFT, what would >> be an application do to another FFT? I have written an Matlab script to >> play around with this idee. When I run the script, the the FFT of an FFT >> gives me a time domain signal, but the result has a larges amplitude. When >> I do a 3rd FFT I get the frequency domain again, and it is also amplified. >> >> It is just something that I picked up. >> >> Heystek >> >> On 23 Apr 2020, at 11:51, James Smith <jsm...@ska.ac.za> wrote: >> >> Hello Heystek, >> >> You can probably use a "to workspace" sink, then you'll be able to >> display the data however you want in some matlab code once the simulation >> is finished running. >> >> Canonically, just applying an FFT to frequency-domain data will get you >> back into the time domain, multiplied by some scaling factor. You need both >> halves of the symmetric FFT though, so the output of e.g. the >> fft_wideband_real wouldn't be meaningful to apply another FFT to it. >> >> Regards, >> James >> >> >> >> >> On Thu, Apr 23, 2020 at 9:40 AM Heystek Grobler <heystekgrob...@gmail.com> >> wrote: >> >>> Good day Casperites >>> >>> I have an interesting question. I am using a FFT in simulink for the >>> use in a spectrometer design. I want to test the output of the FFT by using >>> some kind of scope. Simulink only has spectrum scope, that would be >>> perfect, but the scope does a second FFT on the signal. The other option is >>> a vector scope, but that does not give the result that I am looking for. >>> >>> Does anyone have a suggestion on how I can see the results of the FFT in >>> simulink? >>> >>> Then I have another question. What would be expected if I do a FFT on a >>> FFT? As far as I can figure out, the FFT of the FFT should just be time >>> reversed? >>> >>> Thanks for the help! >>> >>> Heystek >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "casper@lists.berkeley.edu" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to casper+unsubscr...@lists.berkeley.edu. >>> To view this discussion on the web visit >>> https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/62AFF5A0-0980-4FB7-8CB8-C053590F655E%40gmail.com >>> . >>> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "casper@lists.berkeley.edu" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to casper+unsubscr...@lists.berkeley.edu. >> To view this discussion on the web visit >> https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/CAG67D373fmnXJ7Os7jrBGJyTbvRA835okUMO%2B0x8hTT%3DGzyKrg%40mail.gmail.com >> <https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/CAG67D373fmnXJ7Os7jrBGJyTbvRA835okUMO%2B0x8hTT%3DGzyKrg%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "casper@lists.berkeley.edu" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to casper+unsubscr...@lists.berkeley.edu. >> To view this discussion on the web visit >> https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/8064F663-15CE-4C42-B13A-9AC9ABF82703%40gmail.com >> <https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/8064F663-15CE-4C42-B13A-9AC9ABF82703%40gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- > You received this message because you are subscribed to the Google Groups " > casper@lists.berkeley.edu" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to casper+unsubscr...@lists.berkeley.edu. > To view this discussion on the web visit > https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/CAG67D37sAj_7sUdaJuAhRd95Ho8Xchdy0guhdLqhKO3e7YVGjw%40mail.gmail.com > <https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/CAG67D37sAj_7sUdaJuAhRd95Ho8Xchdy0guhdLqhKO3e7YVGjw%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "casper@lists.berkeley.edu" group. To unsubscribe from this group and stop receiving emails from it, send an email to casper+unsubscr...@lists.berkeley.edu. To view this discussion on the web visit https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/CAG4nf73TVOV-PpmF9sS-W9zwomGBCXC6gjZNCDpWhj%3DpT0%3D2Zw%40mail.gmail.com.
