This question already has an answer here:
For example, for an aperiodic gate pulse, the Fourier Transforms for the continuous time case is a sinc function, while the discrete time case gives a sine over sine periodic kind of a function. In either of these cases, we get a spectra that goes below the frequency axis, i.e., acquires a "negative amplitude" (namesake) on many occasions in its excursions. I realize that the spectra gives a measure of "how much" a frequency component is present. But how this negative amplitude satisfies this criteria is what I am missing. Is it kinda like.. "How much less that frequency component is present"..? I know I might not be making any sense. Please help. Thanks!!