Description: Modern methods of managing financial, informational and other resources are carried out using access systems that are based on passwords. These systems are simple and convenient, but have low reliability. This is confirmed by almost daily reports of financial embezzlement. Improving the reliability of access systems associated with the use of static biometric features, such as fingerprints of a user. Unfortunately, these expectations were not met since the latter can be easily faked. Therefore, intensive research is currently being carried on the use of dynamic (behavioral) features of a user and, above all, his voice signal in ac-cess systems. However, voice authentication systems are not very reliable. The aim of this work is to analyze the current state of voice user authentication systems in infocommunication systems and, on its basis, determine the directions for improving the quality characteristics of access systems. The current state of voice systems is analyzed the given paper. Studies have shown that at present, these systems are based on extracting user characteristics from the amplitude-frequency spectrum, and the phase data of the voice signal is ignored. Due to the low informativeness of the features obtained, it is necessary to complicate decision-making procedures. The main attention in the given article is focused on the procedures for the formation and use of the phase information of the voice signal. In the process of experimental studies, it was found that the phase-frequency spectrum is more informative and makes it possible to select one and a half times more formant frequencies with respect to the amplitude-frequency spectrum. The latter will significantly affect the quality of user authentication procedures. Further studies will focus on finding the distinctive features of a user according to the phase information of his voice signal.
Keywords: amplitude, authentication, voice signal, spectrum, frequency, phase.