A Synthetic Voice for an Assistive Conversational Agent: A Survey to Discover Italian Preferences regarding Synthetic Voice’s Gender and Quality Level

Figure 3

Differences among adolescents, young adults, middle-aged participants, and seniors’ assessments in voice quality scores attributed to the high-quality female and male voices and low-quality female and male voices, respectively. The symbol “” above the bars indicates that scores are significantly different. Mean ranges vary between 6 and 30 (where lower scores correspond to positive evaluation, while higher scores reflect negative evaluation), since the total score is obtained by adding the participants answers (from 1 to 5) to 6 questions.