Paper abstract:

In forensic speaker identification process voice samples are compared on the basis of many speech characteristics through which speakers may differ from one another. These characteristics should be chosen on a case-by-case basis, considering the unique features of the search samples.

In the research process, it is very important to determine whether the similarity (or difference) between the compared samples according to a certain parameter is speaker specific or is a common in the relevant population. The more uncommon a characteristic is among speakers, the more effective it will be in differentiating between them. Consequently, it is important for each language to have statistical data on the prevalence of certain speech parameters. The research aims at collecting this statistical information on the Georgian language.

Research parameters can be categorized as segmental and supra-segmental units of speech, voice and manner of speaking. Some signs can be analyzed both auditory and acoustically. The most common characteristics are: language, dialect, accent, speech rate, fundamental frequency, voice quality, formant frequencies, hesitation, speech pathologies, and etc. (ENFSI, 2022). In this work we discuss the speech rate and fundamental frequency for the Georgian-speaking population.

To collect data on the Georgian native speaker population, we selected three age categories - [18-30], [31-40], [41-50]. In each category was recorded 10 women's and 10 men's Unplanned speech in dialogue mode (60 speakers in total). On the obtained audio recordings, we measured the articulation rate based on the number of syllables uttered in a time interval excluding pauses, hesitation elements and disfluencies and the fundamental frequency of the speakers.

Research shows that there is no significant differences in age and sex groups by articulation rate in Georgian language. Average articulation tempo is 5.7 syllables per second and most common tempo ranges between 5.1 to 6.3 syllables per second.

Data analysis shows that the fundamental frequency decreases with age for both men and women. However, the gap between the age categories of 18 to 30 and 31 to 40 is substantially wider than that between 31 to 40 and 41 to 50. From a physiological perspective, the reason for this may be the still-developing body. Additionally, the fundamental frequency is less distributed in female speakers who are between the ages of 41 and 50.

The current study also includes the description of intra-speaker variation ranges in articulation rate and fundamental frequency in Georgian speaker population.