Speaker identification is a part of phonetics which is called forensic phonetics. The problem is to choose such parameters of speech which are the most stable in the individual speech of person.