Forensic phonetics is concerned with resolving disputed content of the recordings, preparing lines for voice recognition of the suspects, the detection and analysis of the similarities and differences of voice samples, which finally allows the identification of the speaker. Such identification is based solely on the voice of a person which is used to express one's individuality, uniqueness and belonging to the human race – as each man is an unrepeatable individual per se, so is his voice. These differences are attributed to different dimensions of speech organs, but also to different regional accents of a certain language. Information obtained on the basis of accents, dialects and other features of speech allow profiling of speakers that contributes to the process of identifying speakers in forensic phonetics. Although the basic assumption is that different speakers have different voices, it is necessary to know that the voice of a single speaker also varies. Variations of speech within the same speaker can be a result of factors and emotional states which are out of control of the speaker, such as a flu, stress, panic, anger; but the human voice can be changed according to circumstances in which the speech is produced. In the forensic identification of speakers, there can happen voice disguises in order for the speakers to falsify or cover their identity. The identification of speakers includes comparison of voice samples from two sources – one is from the evidence material that comes from the offender which is to be identified, and the other one is from the familiar speaker, i.e. a suspect, as well as form a certain number of „false“ voices (foils), and the task of the expert is to note the similarities and the differences between them. Identification of the speakers can be divided into two basic types – the first one is a naive identification where common human capabilities are used; and the second, specialized/technical type where specialized techniques are put in practice. Naive identification, therefore, is conducted by a „common person“, i.e. a listener who is not an expert and who is lacking linguistic and phonetic knowledge and experience. Probably the most important variable of naive identification depends on whether the voice sounds familiar or not to the listener, and is thus split into two categories: those in which the voice of the accused is known to the listener, and those in which this is not the case. In the technical identification the voice recording is studied by an expert in the analysis of speech, as requested by the police and/or prosecutor. In forensic phonetics such an analysis is evaluated as the most reliable procedure to identify the speaker. Just like the naive identification of the speaker is being cathegorized according to the familiarity of the voice, in the same way 52 forensic foneticians classify technical identification as either auditory forensic analysis of voice or acoustic analysis. Foneticians who use auditory analysis listen to the voice and sound in a way as to enable the phonetic transcription, i.e. to enable transcription of someone's pronunciation, of something what a person has actually said on a particular occasion, while acoustic phonetic analysis uses a particular computer program for speech analysis, after which the results are being compared. Computer software such as Praat allow extraction of statistical data relating to the basic frequency of the speaker's voice and measures features such as vowels or consonants form. Data is extracted from the spectrogram where the intensity of the components of the spectrum is expressed by a degree of opacity on paper, so that a darker lines on the hight of paper that shows corresponding stronger spectral frequencies, while brighter lines shows the corresponding lower spectral frequency. The dark lines are actually formants, .i.e. traces of increased acoustic energy and are based on the shape of the mouth cavity and speech organs and exactly them are essential for a comparison and for an identification.