It is interesting to calculate the highest frequency of multiplicative noise that will just begin to affect the spectrum. That is, the frequency which will just begin to mix adjacent spectral elements rather than convolving noise across the entire spectrum. For a rapid-scan measurement with 5 kHz modulation of the laser, it is 1.3 Hz. Not coincidentally, this is the frequency that goes through exactly one cycle during the interferogram measurement time. The same observations are also true for step-scan measurements, but the frequency is 5000 times lower, or 0.24 milliHertz.