Hi,

I'm asking my question here after an inefficient search.

Could someone explain precisely why the lomb normalized periodogram based on the fasper routine gives 4 times more abcisse (wk1) points than the fast fourier transform (considering only useful points of FFT, i.e. n/2), meaning a best spectral resolution?

fasper outputs an approximation to a continuous spectrum. You can get as dense a sampling of this function as you want by setting the variable ofac. Also, since some of your unequally spaced input points may be closely spaced, there is no hard limit to how high in frequency you can look -- this is set by the variable hifac.

HTH.