How to calculate the IFFT of a frequency domain IRF obtained experimentally?

Hi, folks!

I am not expert in DSP, actually, I use it rarely. However, I have found a problem to face quite different than the usual.

The problem is very simple: given a complex signal (obtained experimentally) in the frequency domain that represents a impulse response function (IRF), calculate its representation in time domain. The IRF (magnitude and phase) is illustrated in the graph below.

The data are: frequency (w), magnitude and phase.

Firstly, it seemed as simple as apply IFFT, but it does not. My first surprise was the result of IFFT is a complex number. Unexpected once I expected to have a time domain signal. The other problem would be calculate the time axis, because I do not have any information. I must say that the only information known about the time domain is that the IRF is non-casual.

If someone could help me and give some direction I would be very grateful!

Answer 1

[Moonhee]

Question to Star Stride :)

In your code you have information about the original signal in time domain, and hence the fft ofs. I would like to know how to extend the magnitude and phase values in the frequency range,without any information about fft(s) to at least twice the nyquist frequency, in order to obtain the same time signal. Could you tell me how to do this?