Take Fourier transform of an imat_fn.
g=fft(f)
g=fft(f,AMPUNITS)
FFT converts between time and frequency-domain functions. It automatically determines the FFT conversion direction based on the function type. Time-domain functions are converted to frequency domain, and vice versa.
F is an imat_fn of time and/or frequency domain functions. AMPUNITS is an optional string to override the amplitude units specified in the function.
Valid values for AMPUNITS are
'half-peak' - Default(also known as 0-peak amplitude)
'peak'
'RMS'
When converting from time to frequency domain, only the real part of the abscissa is retained, and only the "positive frequencies" of the result are returned. Thus, the FFT of a 1024-point time history will be a 513-point spectrum. By default, the spectrum will be normalized to half-peak, so that a sine wave of unit amplitude will have a spectral amplitude of 0.5 (this is consistent with I-deas Test). You may override this with the optional AMPUNITS argument.
When converting from frequency to time domain, a real function is returned. The AmplitudeUnits attribute of the spectrum will determine the interpretation ('Unknown' will be treated as 'Half-peak').
>> f=imat_fn(1);
>> f.abscissamin=0;
>> f.abscissainc=1/100;
>> f.ordinate=sin(2*pi*(0:999)/10);
>> setdisplay(imat_fn,'functiontype','amplitudeunits','numberelements');
>> g
g =
IMAT Function with the following attributes:
Record Name FunctionType AmplitudeUnits NumberElements
--------------------------- ---------------- ---------------- ----------------
1_(1X+,1X+) Spectrum Half-peak scale 501
>> max(abs(g.ordinate))
ans =
0.5000
>> h=fft(g)
h =
IMAT Function with the following attributes:
Record Name FunctionType AmplitudeUnits NumberElements
--------------------------- ---------------- ---------------- ----------------
1_(1X+,1X+) Time Response Half-peak scale 1001
>> max(abs(h.ordinate))
ans =
1.0010
>>