A Comparison of an Artificial Intelligence Tool to Fundamental Frequency as an Outcome Measure in People Seeking a More Feminine Voice

Validation of an AI-assisted Treatment Outcome Measure for Gender-Affirming Voice Care: Comparing AI Accuracy to Listener's Perception of Voice Femininity

OBJECTIVES

There is currently a lack of objective treatment outcome measures for transgender individuals undergoing gender-affirming voice care. Recently, Bensoussan et al developed an AI model that is able to generate a voice femininity rating based on a short voice sample provided through a smartphone application. The purpose of this study was to examine the feasibility of using this model as a treatment outcome measure by comparing its performance to human listeners. Additionally, we examined the effect of two different training datasets on the model's accuracy and performance when presented with external data.

METHODS

100 voice recordings from 50 cisgender males and 50 cisgender females were retrospectively collected from patients presenting at a university voice clinic for reasons other than dysphonia. The recordings were evaluated by expert and naïve human listeners, who rated each voice based on how sure they were the voice belonged to a female speaker (% voice femininity [R]). Human ratings were compared to ratings generated by (1) the AI model trained on a high-quality low-quantity dataset (voices from the Perceptual Voice Quality Database) (PVQD model), and (2) the AI model trained on a low-quality high-quantity dataset (voices from the Mozilla Common Voice database) (Mozilla model). Ambiguity scores were calculated as the absolute value of the difference between the rating and certainty (0 or 100%).

RESULTS

Both expert and naïve listeners achieved 100% accuracy in identifying voice gender based on a binary classification (female >50% voice femininity [R]). In comparison, the Mozilla-trained model achieved 92% accuracy and the previously published PVQD model achieved 84% accuracy in determining voice gender (female >50% AI voice femininity). While both AI models correlated with human ratings, the Mozilla-trained model showed a stronger correlation as well as lower overall rating ambiguity than the PVQD-trained model. The Mozilla model also appeared to handle pitch information in a similar way to human raters.

CONCLUSIONS

The AI model predicted voice gender with high accuracy when compared to human listeners and has potential as a useful outcome measure for transgender individuals receiving gender-affirming voice training. The Mozilla-trained model performed better than the PVQD-trained model, indicating that for binary classification tasks, the quantity of data may influence accuracy more than the quality of the data used for training the voice AI models.

Outcomes of Gender-Affirming Voice and Communication Modification Training for Non-binary Individuals: A Case Series

OBJECTIVES

There is currently no research reporting solely on outcomes of voice and communication modification training (VCMT) in individuals who identify as non-binary and genderqueer (NBGQ) in the English literature. This study aimed to describe the objective and subjective impact of VCMT on the voice of NBGQ individuals undergoing a 12-week gender-affirming VCMT program.

METHODS

A retrospective consecutive case series of NBGQ individuals enrolled in a VCMT program was performed. Demographics, Transgender Self-Evaluation Questionnaire (TSEQ), fundamental frequency (F0), and frequency range were collected before and after the program.

RESULTS

Four NBGQ individuals enrolled between January 2019 and June 2021; the mean age was 27.0 years. While all four participants represented in this case series showed improvement in at least one of their initial goals, only one improved both their F0 and TSEQ scores; the other three participants had mixed results.

CONCLUSION

NBGQ individuals experienced improvements in self-reported outcomes and changes in acoustic measures after completing VCMT in our case series. Individuals experienced significant improvement in subjective outcomes despite small changes in acoustic measures, and vice versa. More research is needed to better understand the voice and communication needs of NBGQ individuals, along with their outcomes with VCMT.

LEVEL OF EVIDENCE

Level 4.