< Mujpy.Fitplot | Index | Mujpy.FFTResidues >
Only performed in mufitplot result graphics. Fit is always performed on original, rebinned data.
- asymmetry data and errors are mixed with a reference at {$\nu_{\mathrm R}$}, i.e. multiplied by {$2\cos(2\pi\nu_{\mathrm R}t)$}, so that each original real {$\cos\nu t+\phi$} Fourier component of the asymmetry gives rise to two additive terms,
{$\cos(2\pi(\nu-\nu_{\mathrm R})t+\phi)$} and {$\cos(2\pi(\nu+\nu_{\mathrm R})t+\phi)$} and
- the high frequency term, at {$\nu+\nu_{\mathrm R}$}, is filtered by masking high frequencies in the real fast Fourier transform rfft of the mixed signa,l before the inverse operation, irfft.
- the plot should be done with an intermediate packing factor, such as to affect moderately the lower frequency term and further suppress the high frequency residuals.
- Because of the double FFT transformation, if the original asymmetry is over the time interval {$(t_0,t_1)$}, the reconstructed data {$A_{\mathrm R}$} suffers from the unphysical constraint
{$A_{\mathrm R}(t_0)=A_{\mathrm R}(t_1)$}
For this reason apodization is obtained by zero padding the data up to {$t=2t_1-t_0$} prior to FFT
< Mujpy.Fitplot | Index | Mujpy.FFTResidues >
