A Study on the Methods and Algorithms Used for the Separation of Speech Signals

Despite the fact that voice recognition is widely used and has many advantages, there are also a number of serious disadvantages of the methods used in it. This includes, first of all, the weakness of the ability to distinguish between the methods and the number of errors of the first type, that is, the number of falsely denied access, and the second type of error is the most dangerous, that is, the number of people falsely admitted to confidential information, who do not have access. This is especially complicated when performing identification in real conditions accompanied by a set of unfavorable external factors. In this work, methods and algorithms for distinguishing signs of speech signals were studied, and the results of experiments were obtained on the basis of the “rabbit-filter” speech generation model and the algorithm for finding Mel-cepstral coefficients.

Ҳали таржима қилинмаган