Phase analysis of human speech for persons recognition through voice

摘要

The method of approximation is used for decomposition of different sounds of human speech on modes. The data on five specific sounds received from 11 respondents are presented. The behavior of the amplitudes and phases of the modes of sounds is examined. It was found that the phases of different modes of every sound are not independent random variables, on the contrary, there is dependence between them. This dependence is unique for each respondent. It indicates to possibility to automatically identify of a person through voice.