On a Monday morning in a closet-sized soundproof room in the Robert F. Pierce Speech-Language-Hearing Clinic on the University of Minnesota Duluth (UMD) campus, Janet Nindorf reads from a computer screen into a headset microphone. 

“Mary may run on the road.”

“Where were you a year ago?”

“Jo brought wood.”

The sentences—lines from Little Women, White Fang, and other classics—aren’t meant to tell a story or even make sense. Rather, they’re the content of recordings that will be sent to the Nemours/Alfred I. duPont Hospital for Children in Wilmington, Delaware. There, the sentences will be broken into sound particles that will become the raw material for a synthesized voice that will sound like Nindorf’s rather than Siri’s or Alexa’s.

Recording and deconstructing words, then building a voice out of the sounds, is the mechanism behind voice banking, a new technology that Nindorf, who was diagnosed with amyotrophic lateral sclerosis (ALS) in January, is using to hang on to one important piece of her personality.

“I knew my voice was going, and it’s important to be able to connect with people,” the former registered nurse says of her reason for banking her voice. Nindorf traveled from her home in Superior, Wisconsin, to UMD once a week for three months over the summer to record nearly 1,100 sentences, along with names of family members and messages such as “I  love you.”

She is one of 15 people who’ve banked their voices at UMD since March 2017. Jolene Hyppa Martin, who runs the UMD program, says most came from throughout northeastern Minnesota, sometimes driving hours to get to the clinic for one- to two-hour sessions. 

“One woman drove two hours and didn’t speak the entire time in order to preserve her voice,” says Hyppa Martin, a speech-language pathologist and assistant professor of communication sciences and disorders at UMD. 

Nearly all of Hyppa Martin’s clients have ALS, a progressive disease that attacks nerve cells in the brain and spinal cord that control muscles. It affects as many as 20,000 people in the United States at any given time, according to the ALS Association. 

“Most people are diagnosed in their late 50s or early 60s—a time in their life when they’re at the peak of their career, they’re thinking about retirement, they’re starting to have grandchildren—and the bottom falls out with this diagnosis,” she says. 

Studies have found between 80 percent and 95 percent of people with ALS eventually experience speech difficulties. For some, like Nindorf, the voice is the first to be affected. 

When speech-generating devices became available in the 1980s, the “voices” sounded robotic (think of Stephen Hawking). Since then, more natural-sounding synthetic voices have been developed. But they’re expensive, and they don’t have the speaker’s tone, texture, and timbre.

Hyppa Martin explains that voice banking is a work in progress. “The technology is changing so fast. The way we voice bank in 2019 may be very different from the way we do it in 2018,” she says.

SILENT NEED

Rebecca Lulai, director of clinical programs in speech-language pathology at the U of M Twin Cities, has seen voice banking evolve since 2015, when the U’s Julia M. Davis Speech-Language-Hearing Center became one of the first in the country to offer the service. “There was a lot of troubleshooting, rerecording, and working through the kinks,” she recalls of the early sessions. The software interface is now more user-friendly, and clients can make a custom bank of names, words, and messages that can be “read” by a speech-generating device.

Lulai has worked with clinicians from UMD, St. Cloud State University, Minnesota State University  Moorhead, and Minnesota State University, Mankato to start voice banking programs in other parts of the state. 

Hyppa Martin, who earned her doctorate at the U and who works with people with severe communication disorders, learned about voice banking through the ALS Association. “I decided it was something I wanted to offer at UMD,” she says. 

She initially expected one request for the service each semester but has seen quadruple that number. Hyppa Martin not only works with clients, but also is advancing the field. “There isn’t much information out there on voice banking,” she says. “When I did a systematic search, I found only three peer-reviewed articles on the topic.” 

She has been collaborating with Lulai, clinicians from St. Cloud State, and students from UMD to share their collective knowledge at speech and hearing conferences and help others start voice banking programs.

Since insurance doesn’t cover voice banking, the ALS Association and the Gleason Initiative Foundation now pay for the software used to make and process the recordings and an app that can turn an iPad into a speech generator. At UMD, Hyppa Martin volunteers her time to help clients. 

Because of support from individuals and the Edwin H. Eddy Family Foundation, the UMD clinic is able to offer this and other services at no charge. “We’re on a shoestring budget, and we’re constantly trying to stretch every dollar,” she says. “To a certain extent, philanthropy has kept our doors open.”  

BUILDING A BETTER VOICE 

Gifts from donors also have provided scholarships to students who are improving voice banking. Claire Barnes, a senior studying speech-language pathology, received several scholarships, including one established by Evelyn Anderson, ’55 B.S., specifically for a student involved in voice banking (see “Banking on the future”). 

Barnes and graduate student Jaclyn Friese, ’18 B.A, created a how-to manual based on lessons they’ve learned at the UMD clinic (use a foam microphone tip if the client sounds breathy, for example, and avoid wearing clothing and jewelry that can make sounds during recording sessions).

They also built a prototype of a portable soundproof booth that could allow people to record their voices at home. “Without people’s philanthropy, I never would have been able to do any research in the first place,” says Barnes, who plans to attend graduate school. “And I wouldn’t have been able to do these projects without the Anderson scholarship.”

Meaghan O’Connor, a senior in the communication sciences and disorders program, has been looking into people’s preferences regarding voice banking. Earlier this summer, she surveyed Native American elders at a health fair about whether they would prefer a computerized voice that’s easy to understand or a voice that sounds like their own but may not be as clear. 

At the Minnesota State Fair, O’Connor worked with Hyppa Martin on a research project where they played short recordings of computer-generated voices and asked participants to rate the quality. 

O’Connor says scholarships from the Duluth Lions Club and others afforded her the time to do this work, which is critical to getting into graduate school. “It’s very competitive to get into graduate programs for speech-language pathology,” she says. “Now I’ll be able to show that I have experience with research projects.”

Hyppa Martin, who has exposed about 25 students to voice banking, will continue working with them to improve the process and the end product. She would like to lessen the time and effort needed to bank one’s voice, make synthetic voices sound more natural, and start an outreach program so people can record their voices at home. “It’s a win-win if I can serve the community, engage in research, and bring in students,” she says. 

But as much as she enjoys working with voice banking clients, Hyppa Martin hopes she one day won’t have to. “We need a cure for ALS,” she says. “But until we have one, this is one contribution I can make to people facing this condition that can improve their quality of life.”

Kim Kiser is editor of Legacy magazine.