el_supremo:
Why do you "need to perform" the IFFT? What is it going to tell you?
are you for real ?
FFT only is virtualy useless unless you want to display a spectrum on an lcd or something
IFFT bring the whole power of using fourier transform
FFT only is virtualy useless
For some things, like changing the oil in your vehicle, sharpening a pocket knife and crushing gravel.
However, the FFT algorithm is very heavily used for other purposes.
IFFT bring the whole power of using fourier transform
OK, I'll bite. What does the IFFT "bring" that the FFT on its own does not?
Pete
the ability to manipulate in frequency space, not just observe in frequency space.
in audio work it is crucial. hence why i am trying to find a reliable way to do inverse FFT. im curious about using the CMSIS libraries. its how i got onto this thread
Better to start your own thread, not camp-on to an old one, describing your particular problem in the subject line.
pepelevamp:
the ability to manipulate in frequency space, not just observe in frequency space.
in audio work it is crucial. hence why i am trying to find a reliable way to do inverse FFT. im curious about using the CMSIS libraries. its how i got onto this thread
Using the FFT to manipulate the spectrum is not a good idea.
The spectrum of the FFT as a bank of band pass filters has the shape of sin(x)/x. (It's basically a bank of moving average filters, shifted in the frequency domain.) This means that it has very poor band-pass characteristics, it has pretty high bumps outside of the pass band, so it will interfere with other frequencies.
You can improve this by using a window function before applying the FFT.
https://www.dsprelated.com/freebooks/sasp/DFT_Filter_Bank.html
Depending on the application, you'll be better off using a bank of band-pass filters that were specifically designed for that purpose.
Pieter