Please can someone explain the benefits of a FIR filter with complex coefficients, as opposed to one which uses real coefficients? Why would I choose a complex FIR over a real one? I have seen complex FIR filters used occasionally, for example in equalization, but I've never understood why you can't use real coefficients here.
Also, how would one design a complex FIR filter in Matlab? As far as I know from my limited experience, the Matlab filter designer (sptool or fdatool) only outputs designs with real coefficients.
Complex filter coefficients -- can be required only under special cases.
If the input is complex, then probably one of the following is true
1) The spectrum of the input is not symmetrical
2) There is only one sided spectrum.
You may need special frequency response from the filters [non symmetrical]. Then again you use it.
However my belief is you can do everything using real filters. I donot see where this doesnot work. COmplex coefficients may sometimes simplify the structure, so the fixed point may be more optimized.
As you know, An if signal can be presented by its base band equivalent signal (According to Prokis) that is complex. When you sample a signal with fs as sampling rate, you can study the frequencies between (0,fs/2), but with a base band equivalent signal, you can study (0,fs)!!!!
I think I have seen some new features in matlab 7.4 on designing filter with complex coef.