Interactions between voice cues in the perception of transgender and cisgender voices

Acoustic voice cues, like fundamental frequency (f0) and vowel space area (VSA), are used by listeners to attribute gender to a talker. However, the variability of f0 and VSA in transgender and cisgender talkers makes it difficult to understand how listeners use these cues to make gender judgments. The goal of this study was to determine how listeners use f0 and VSA when making voice gender judgments in transgender and cisgender talkers. Forty-two cisgender listeners used nine-point scales to assess the perceived gender (“definitely male” to “definitely female”) of 30 transgender (man, woman, non-binary) and 30 cisgender (man, woman) talkers. Acoustic measurements of f0 and VSA were computed for each talker, and linear mixed effects models predicted listener judgments from f0, VSA, talker gender identity, and interactions. Results suggest that VSA is an important cue for gender judgments, but that its weighting varies based on f0 and listeners’ perception of gender prototypicality. For transgender talkers with non-prototypical f0 values (e.g., low f0 for trans women), listener judgments are highly dependent on VSA. Results reinforce that there is not a straightforward mapping between the selected voice acoustics and perceived voice gender.