Voice quality control techniques for shaping perceived gender impressions

In recent years, growing societal interest in gender diversity has brought attention to acoustic factors in gender impression formation. This study examined how gender impressions change before and after modifying the fundamental frequency (f o) so that the average f o of each utterance was normalized to a common reference value, based on a subjective evaluation using speech samples in Japanese, English, and Thai. Ten Japanese listeners in their twenties participated in the subjective experiment. The results showed that male voices with lower f o were perceived as more masculine, while female voices with higher f o were perceived as more feminine, indicating that f o has a substantial influence on gender impressions. In addition, the findings suggest that impressions are also affected by other acoustic features such as prosody, as well as by the listener’s native language and cultural background. Therefore, to shape gender impressions as intended, it is necessary to consider the combined effects of these multiple acoustic features. Building on these findings, this study aims to quantitatively analyze multiple acoustic features, including f o and prosodic characteristics, and model them as controllable parameters. The ultimate goal is to establish a voice quality control method capable of producing specific subjective gender impressions for listeners.