A noise signal is a set of simultaneously existing electrical oscillations, the frequencies and amplitudes of which are random. As a rule, it is impossible to trace any regularity in the change in the instantaneous values of the noise. At the same time, such signals have certain probabilistic characteristics. The authors set themselves the task of assessing the quality of acoustic noise generators using various methods: determining the noise quality factor using asymmetry and kurtosis coefficients; determination of the entropy quality factor; determination the entropy coefficient of the quality of the distribution of the envelope of the noise signal. To assess the quality of the generators, an acoustic noise generator, which is used to protect acoustic information in the office, was taken as a control sample. It is also proposed to use a generator created with Python software using a pseudo-random sequence. Based on the results of the work, a conclusion was made about the use of the proposed methods for assessing the quality of noise signal generators.