Background: Auditory-perceptual voice evaluation is essential in assessing voice disorders but faces challenges in reliability. Existing databases for training and research are primarily composed of voice samples evaluated by a small number of raters from limited geographical regions. Given the subjective nature of perceptual evaluations, increasing the number of raters and diversifying their origins is crucial to enhance the reliability and generalizability of these assessments.

Objective: The aim of this project is to begin building a global database of perceptual voice evaluations, incorporating input from users worldwide.

Method: Participants in this study are users of the All-Voiced app (www.all-voiced.com), designed for training and research in auditory-perceptual voice evaluation. Upon registration, users provide demographic data (age, gender, profession, voice experience, country, and primary language) and give informed consent. A selection of 120 voice samples from the Perceptual Voice Qualities Database (Walden, 2022) was made, ensuring representation across age groups, sex, diagnosis, and severity of dysphonia. All samples were normalized to 70dB, and demographic data were standardized for consistency. The app uses various instruments, including the Consensus Auditory-Perceptual Evaluation of Voice (Kempster et al., 2009), to evaluate voice samples remotely. Perceptual ratings are anonymized and automatically collected with sociodemographic data, which are analyzed to explore trends and variability based on participants' profiles.

Results: Preliminary data collected over eight months, from the launch of the latest version of the app on September 4, 2024, is presented. These results include the number of evaluations, user demographics, and comparisons between global user ratings and expert ratings from the PVQD. Additionally, the data compares evaluation results by expertise level and other demographic variables such as age, primary language, and country.

Conclusions: The preliminary findings are promising, showing how global participation, facilitated by technology, diversifies and enriches the PVQD’s data, broadening its utility for researchers, teachers, students, and clinicians. As more data is collected, this growing resource will enhance the reliability and generalizability of voice evaluations across populations, ultimately supporting a deeper understanding of perceptual voice evaluation and enabling the design of better training programs.