“…Regarding the different steps constituting the analysis procedure, Grard et al [2011] state: “The first three steps, though debatable, might possibly help emphasizing, in a qualitative way, spectral components so long as they exist… Adding zeroes to the time series at step (iv) does not contribute any information at all; it merely smooth out the spectrum and provide an interpolation that is not justified, from a mathematical point of view, ….” Contrary to this last statement, the fourth step is the one with a more solid mathematical background; specifically, in Shannon's theorems, also known as sampling theorems. The frequency sampling theorem [ Brigham , 1988, p.86] states that a time‐limited function g ( t ), that is g ( t ) = 0 for ∣ t ∣ > t max , is uniquely determined from the equidistant samples of its Fourier transform, G ( f ), whenever the step frequency Δ f fulfills the condition Δ f ≤ (2 t max ) −1 .…”