A signal processing method of (amongst other things) estimating fundamental frequency from the speech pressure waveform. The word was made up by reversing the first four letters of the word spectrum. The technique involves treating the Fourier transform (essentially the amplitude-frequency spectrum) of the speech pressure waveform as a signal itself and producing a second Fourier transform of that. The resulting representation shows power as a function of quefrency (the cepstral equivalent of frequency). The maximum peak in this cepstral representation corresponds to the fundamental frequency of the original input.
Much more detail is given here.