American telephone voice: Socio-phonetic features of identity in professional and non-professional discourse

Abstract

The study is concerned with the role of socio-phonetic characteristics of American speakers’ voices in identifying gender and age identity in professional and non-professional communication. We have to develop and test the methodology previously applied in such specific areas as corpus analysis, forensic phonetics, multi-modal research, and intraspeaker variability to address the problem of identifying personality by voice. The aim of the research is to view variations in the prosodic forms of American English speakers and examine age-related and gender-specific prosodic features in telephone communication. The study intends to address the following research questions: How do the following prosodic features contribute to distinguishing age- and gender-related changes in the human voice: mean pitch, maximum pitch, minimum pitch, pitch range, mean intensity, jitter, shimmer and harmonics-to-noise ratio? Which combinations of features are associated with young, middle-aged, and senior voices of men and how much are they different from or similar to women’s voices? The auditory and acoustic computer analyses were conducted on American English dialogues with 30 speakers, equally balanced for gender (5 men and 5 women in each age group) and three age groups (young, middle-aged, senior), taken from the American telephone speech corpus to measure the prosodic parameters of pitch, pitch range, intensity, jitter, shimmer, and harmonics-to-noise ratio supported by statistical data processing in Minitab programme. The apparent-time technique of data presentation and the comparative analysis allow discovery of the dynamics of voice changes over time which could be later applied to personality identification. Comparing the data based on eight parameters’ values in three age groups balanced for gender we have found that although all the selected measurements proved to be relevant for either gender group, the age-related trajectory of voice dynamics may be different in male and female voices regarding their specific prosodic characteristics. Prosodic features of mean pitch, maximum pitch and minimum pitch have been confirmed to define both individuals and groups of people of a certain age and gender and could be considered as both speaker-identifying and group-identifying characteristics. Voice quality features, apart from being previously found to signify the emotional states of speakers, may also be observed to characterise certain age and gender groups; hence, they may also serve as speaker-identifying characteristics.

Keywords

• telephone voice
• prosody
• age
• gender
• american english
• socio-phonetic features