Documentation

Flag to specify complex output, same ordering as fftw output. For complex data the ordering is the same as Numerical recipes, fig 12.2.2. For real data the ordering is as follows: if there are n frequency bins labelled from 0 -> (n-1), and r,i represent the imaginary parts respectively, then the ordering is r0 r1 r2 ... r(n-1) i0 i(n-1) i(n-2) ... i1. This funny ordering is a consequence of using a shorter fft to exploit the symmetries of real input data. The units are amplitude rms per root Hz for both real and imaginary parts. This option is the fastest because no rearranging of the output data is required. Ordering is such that, when amplitudes from both positive and negative frequencies are combined, the overall amplitude of the sum in quadrature of real and imaginary parts is the rms spectral density of a sine wave in that frequency bin in amplitude per rtHz.