A distinctive experiential learning assignment saw Brock Applied Linguistics students donate their voices to learn about the mechanics of speaking as well as technological innovations to support the development of synthetic voices.

Students of “Anatomy and Physiology of Speech and Swallowing” auditioned to donate their voices to VocaliD’s Human Voicebank, which uses artificial intelligence (AI) to blend donated voices into bespoke voices for people who rely on synthetic speech tools.

The experience helped students learn about the potential of assisted communication technology as well as the airflow, bones and muscles used while speaking.

“Prior to my voice donation, I was unaware of the insufficient diversity of synthetic voices for those who rely on assistive technology,” says Hailey Gray, a fourth-year Child Health student minoring in Applied Linguistics. “After reading a TED Talk by Rupal Patel as part of my donation, I realized that a person’s voice is a reflection of who they are in terms of their age, size and personality, so a synthetic voice should reflect the same.”

Once approved as donors, students logged up to 10 hours over the course of the Fall Term by reading for several 30- to 40-minute sessions — a task that brought a keen awareness of how the voice is typically produced and how the body parts involved become taxed.

“We learnt in class about good vocal hygiene and straining our voice,” says second-year Speech and Hearing Sciences student Jenna Angle. “At times, I read for over 40 minutes and noticed my throat starting to hurt, so it was important to take breaks and drink lots of water to help protect my voice.”

Hilary deBoer, a non-degree student taking courses to prepare for a graduate program in speech-language pathology, says that the experience of being a “surrogate” talker brought the theories studied in class to life.

She explains how the voice donation project broadened her understanding about source-filter theory, which covers how source — sound generated by the larynx — works with filter — the rest of the vocal tract that shapes sound — to create speech.

“Using the voice donations, the technologists can combine the source of someone who’s not able to articulate speech on their own with as little as one recorded vowel sound and combine it with someone else’s articulation and resonance,” says deBoer. “It creates a synthetic or prosthetic voice that is unique to the client.”

Instructor Charlene Cratt, a former Speech-Language Pathologist who is passionate about advocating for people who use augmentative and alternative communication, says the assignment ticks several boxes in terms of what she wants to impart to students.

“There’s the professional angle in terms of opportunities for work and career paths in technology, there’s the personal angle of reflecting on the experience of donation and there’s the application of the physiology itself,” says Cratt. “I have done this assignment before with students, but the magnitude of the donation this year really was impressive, with over 100 hours of digital speech donated by the class and each participant donating at least 1,500 sentences.”