"Obviously, the channel separation ??f cannot be very small unless each of the frequency spectrums f1, f2 . . . fN is very stable, that is, if those frequencies do not change appreciably as a function of time, temperature, or other factors." . . . . .