学术文献库

The present study aims to explore user acceptance and perceptions toward different quality levels of synthetical voices. To achieve this, four voices have been exploited considering two main factors: the quality of the voices (low vs high) and their gender (male and female). 186 volunteers were recruited and subsequently allocated into four groups of different ages respec-tively, adolescents, young adults, middle-aged and seniors. After having randomly listened to each voice, participants were asked to fill the Virtual Agent Voice Acceptance Questionnaire (VAVAQ). Outcomes show that the two higher quality voices of Antonio and Giulia were more appreciated than the low-quality voices of Edoardo and Clara by the whole sample in terms of pragmatic, hedonic and attractiveness qualities attributed to the voices. Concerning preferences towards differently aged voices, it clearly appeared that they varied according to participants age' ranges examined. Furthermore, in terms of suitability to perform different tasks, participants considered Antonio and Giulia equally adapt for healthcare and front office jobs. Antonio was also judged to be significantly more qualified to accomplish protection and security tasks, while Edoardo was classified as the absolute least skilled in conducting household chores.