I use a combination of Visual Basic Programming interfaced with FAMOS which is a DSP software like MATLAB but much easier. I am trying to separate signals from a group, which requires filtering each signal. I used FFT and then IFFT on a 600s signal but the IFFT was longer than the original. I don't understand if this changes the originality of the signal in terms of shape,which might affect the further processing of the signal. After re-sampling the original 10min signal according to Shannon-Niquist theorem and using rectangular windowing, I detect the frequency using FFT.

  1. After I apply IFFT to the signal, does this change the phase or amplitude even slightly that might vary from the original signal?

  2. How do I filter one particular freq before IFFT after detecting it from FFT (Maybe I use peak detection)? Is there a general method of doing it? DSP is totally new to me,and I am still attempting to learn.


Frequency domain processing (i.e FFT, manipulating signal, IFFT) is actually pretty complicated and requires careful choice and a thorough understanding of things like framing, step size, windowing, overlap, wrap-around, circular vs. linear etc. Simple examples for these type of algorithms are overlap-add or overlap-save for linear time invariant filtering.

If the only thing you want to do is getting rid of the low frequencies in your signal, filtering with a high pass filter is much easier and quicker.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy