Characterization of Sub-Nyquist signals with analog modulation, based on the distribution of singular values The work evidences the probability distributions that present the singular values, obtained from the autocorrelation matrix of the continuous signal samples, acquired with Sub-Nyquist random demodulator sampling. The results come from an experimental process at simulation level developed in Matlab®. The simulation process starts from the experimentation with pure sinusoidal tones, continuing with signals with a.m. modulation, and ending with frequency modulation. These signals are immersed in Gaussian additive white noise, in order to get closer to reality. Key words: Autocorrelation, AWGN, spectrum, sampling frequency, cognitive radio, Sub-Nyquist, singular values.