Age and gender preferences for synthetic and natural speech

Abstract

The purpose of this investigation was to examine the preferences of younger and older male and female listeners with regard to natural and computer-generated synthetic speech in six different contexts. The subjects (listeners) for this study included five males and five females in each of four age groups (6- to 8-year-olds, 10- to 12-year-olds, adolescents, and adults). The listeners rated their preferences for 12 different voices (four natural and eight synthetic) on a 5 point Likert scale. Their preferences were rated for six communication contexts dependent on the potential user of the voice (adult male, adult female, child male, child female, computer, and self). The data were analyzed separately for each of the six communication contexts. The results indicated that the Smoothtalker 3.0 male and RealVoice female voices were the most highly rated of the synthetic voices. However, there were almost always statistically significant discrepancies between the ratings for natural and synthetic speech. The results of this investigation raise a number of issues related to the combined effects of age and gender-appropriateness of natural and synthetic speech. These are discussed in terms of their implications for the future development of synthetic speech technology used in communication devices.

• age
• attitudes
• gender
• natural speech
• speech synthesis
• technology
• VOCA