• Victoria A. Brownworth

Restoring Lost Voices

Syrinx is an AI-driven wearable device that returns human speech and, with it, human dignity.

Spring 2022 Stanford Social Innovation Review

Syrinx is a 3D-printed electrolarynx device that is worn around the neck and creates vibrations to replicate speech. (Photo courtesy of Syrinx)

Speech is a right, not a privilege” is innovators’ mission guiding Syrinx, a new electrolarynx (EL) device that uses machine learning to recreate speech and vocal patterns to return voices to people.

Syrinx was invented by University of Tokyo engineering graduate student Masaki Takeuchi. He participated in a university civil society program to address problems faced by people with disabilities, and with a team of fellow students—Kunhak Lee, Jaesol Ahn, Yuki Ogasawara, Mizuki Araki, Yutaro Soejima, and Karin Kiho—designed and built Syrinx.

In human speech, air from the lungs moves to the windpipe and vibrates the vocal cords. The vibrations, in turn, carry to the mouth, where the lips form syllables. But when a person’s voice box is removed, there is nothing to vibrate.

Takeuchi had watched a YouTube video featuring a person who had a laryngectomy, a medical procedure that removes part or all of the voice box, rendering that person speechless. To date, the only voice replacement systems have been bulky devices that are handheld, sticklike contraptions placed against the stoma—the opening in the windpipe that remains after a laryngectomy.

“The laryngectomee [in the video] spoke in esophageal speech,” Takeuchi recalls, “but the voice quality was very poor.” The video inspired him to explore voice loss and how it affected the tens of thousands of laryngectomees each year—nearly 12,000 in the United States annually. Worldwide, 300,000 people lose their voice every year due to cancer. “I went to the Ginreikai, a Japanese laryngectomee community, and interviewed them about what they needed,” Takeuchi says. “They said, ‘We want to speak in our natural voice in public.’ So, we made the hands-free wearable electrolarynx named Syrinx, which restores the voice of laryngectomees.”

Takeuchi designed the initial prototype for Syrinx in August 2019 and the current prototype, which improved the quality of human voice reproduction, in March 2020. The mechanics of his design borrow from nature: common hill mynahs—talking birds residing in South and Southeast Asia—have two voice boxes, and they mimic human voice by mixing the sound produced by each voice box.

Syrinx is 3D printed and worn around the neck like a collar. It works by manually vibrating the throat to create sound that a laryngectomee then forms into words using their mouth and tongue. People are able to speak in their natural voice by moving their mouth—meaning that the sound of their voice does not have the robotic tone of traditional EL devices. While handheld devices only approximate vocal sounds because they have limited wave patterns, Syrinx “uses more complex vibration patterns generated from AI using real voices,” Takeuchi says.

People in the Ginreikai community who tried the Syrinx prototype were impressed by the sound. While using the device, Takashi, a laryngectomee patient, said, “I can use both hands when speaking. I want to wear it and talk in public!”

Team Syrinx (formerly called NUTONE) were winners of the 2020 Microsoft Imagine Cup Asia Regional Final. Takeuchi and his team also won the James Dyson Award and the Japan Healthcare Business Contest 2021 Grand Prix Prize. Cash awards from such prizes—$8,000 from the Microsoft Imagine Cup and $2,800 from Dyson—have provided critical funding for Syrinx. Takeuchi also relies on his graduate program scholarship ($1,800 per month), in addition to more recent grants from Japan’s Ministry of Internal Affairs and Communication’s 2021 INNO-vation Disruptive Challenge ($30,000) and Japan’s New Energy and Industrial Technology Development Organization ($50,000).

“I have grants, but we need much more,” Takeuchi acknowledges. “Syrinx has not been sold yet. It is still in prototype.” The goal, he says, is to sell Syrinx within three years. “We have to solve some issues,” Takeuchi explains, “such as adding a pitch, or intonation, and inhibition of the vibration sound.” More funding will be necessary, too, to mass-produce Syrinx in order “to do a larger user test and sell it overseas,” he adds.

Team Syrinx is dedicated to realizing its mission of improving the lives of people who have lost their voice due to cancer, disease, or trauma. “I want to revive people’s lost abilities and identity in order to realize a world where their individuality and personality are respected,” Takeuchi says.

9 views0 comments