Abstract Digital Auto- and Cross-spectral Analysis has become an important method for quantifying and analyzing electroencephalographic data. This paper presents a set of four procedures written in ALGOL, which perform spectrum analysis of up to 14 simultaneously recorded EEG channels on a medium to large size computer. The first procedure reduces the data after smoothing by a digital low pass filter with variable cutoff frequency. By the elimination of the higher frequency components of no interest, aliasing is prevented. The reduced but due to multiplexing still interleaved data are then rearranged by channels. The third procedure performs the Fast Fourier Transformation. Finally, the spectral quantities as power spectrum, cross-spectral amplitude, phase, coherence and gain are computed for preselected channel combinations. Smoothing in the frequency domain is performed either by a rectangular or a truncated normal spectral window of variable width.