Every so often, Robert White breaks into song and serenades his 13-year-old son Kieran with a tender Irish lullaby, “That Little Boy of Mine,” taught to him by his father.
But the deep-throated voice that used to fill his family’s living room in West St. Paul now quivers and shakes. There are times when White, who has amyotrophic lateral sclerosis, or ALS, a terminal disease that impairs his motor skills, can barely finish the chorus before his voice dips to a hoarsened whisper.
For White, and thousands of other Minnesotans afflicted with neurodegenerative diseases, losing one’s ability to speak is one of the most devastating consequences.
Now, however, researchers have found a way to preserve the unique essence of a human voice — in all its idiosyncratic nuance and power — for people with incurable and often debilitating illnesses.
Using new voice database technology, University of Minnesota speech pathologists can record people saying hundreds of sentences and phrases, break them down into phonetic units, and then reconstruct a personalized voice that can be used on a speech-generating device. The end result of this process, known as “voice banking,” is a voice that is nearly identical to the person’s original, healthy voice.
While the technology is still new, many people feel they are in a race against time to store their original voices before they become unrecognizable to their families and friends. Even the simple act of saying “I love you” can be too much of a strain for people in the later stages of ALS or other neurodegenerative illnesses.
“This is about preserving a person’s essential dignity, rather than having to depend on a canned synthetic voice that many find dehumanizing,” said Dr. H. Timothy Bunnell, director of the speech-language laboratory in Wilmington, Del., that pioneered the voice-banking technology about a decade ago.
But the process of preserving a voice is an emotional one for families. It comes with the recognition that a loved one is dying, and that even a carefully reconstructed voice — built over a period of weeks or months — will never be an exact match to a person’s original speech. Even the most sophisticated audio technology cannot recapture the spontaneous emotion of someone who bursts into laughter, or the gentle timbre of a parent’s voice as they sing to a child, researchers say.
On a recent morning, Wendy Eickhoff, 49, struggled to contain her emotions as she began the arduous process of recording more than 1,600 preselected sentences at a speech laboratory tucked in a brick building on the University of Minnesota’s East Bank campus.
Last October, Eickhoff was diagnosed with a rare form of ALS known as bulbar onset, which first attacks the muscles that move the tongue, mouth and vocal cords. It was a particularly brutal diagnosis for Eickhoff, a self-described “talker” who works as a technology relationship manager at Wells Fargo. Slurred speech and weakened tongue muscles were among her first symptoms.
“It was devastating because I have always relished talking,” she said.
As Eickhoff rattled off sentences in a soundproof room, including lines from the “The Velveteen Rabbit” and “The Wizard of Oz,” her daughter Logan, 21, watched intensely through a window in an adjoining room. With a tinge of sadness, Logan expressed hope that voice banking would preserve the unique character of her mother’s speech, from her high-pitched laugh to the gentle way she called out “Good morning sunshine!” to her daughter each morning.
At one point, Logan broke down in tears after a speech pathologist calmly informed her that her mother’s recorded voice would survive her death and would always be available to the family.
“It’s difficult, really difficult,” said Logan, who is pursuing a degree in linguistics. “I don’t want to have to acknowledge that in the potentially near future that my mom won’t be able to speak. But I also feel lucky, because this means that even after she’s gone, I’ll be able to hear her voice. It’s bittersweet.”
Those who have been through the process speak of it in glowing terms.
When Robert White finished recording the last of 1,610 sentences last month, after five grueling sessions, he threw up his hands in victory and handed out assorted chocolates to the speech-language pathologists and young students who helped him through the process.
“There is a real sense of accomplishment when you get to the end,” said his wife, Argerie White. “I mean, how wonderful would it be, if Bob ever loses his voice, that we can still hear him sing?”
Until recently, people like the Whites and the Eickhoffs relied on interactive whiteboards or robotic synthesizers, similar to that used by renowned physicist Stephen Hawking, to communicate with the world after illnesses had taken their speech. But voice-banking technology has become so sophisticated that it can now cull through thousands of phonetic syllables and sort them in a way that replicates the pattern of a person’s speech. The recording process emphasizes the repetition of vowels, because the medley of “ohs” and “ahs” are largely what make a person’s voice distinctive, researchers say.
While the technology is still being tested and enhanced, that has not dissuaded the regional chapter of the ALS Association from encouraging people with the disease to take advantage of the technology while their voices are still strong. The voice-banking service is now available at university speech clinics throughout the Upper Midwest, from Fargo to Duluth. The cost of creating a voice can run from a few hundred dollars to more than $1,000, depending on the provider. In Minnesota and the Dakotas, the cost for those with ALS is covered by the regional ALS Association.
“It’s crazy-amazing how well this works,” said Rebecca Lulai, a clinical supervisor and speech-language pathologist at the University of Minnesota.
Pete Klinkhammer, who remembers the exact minute he was diagnosed with ALS (11:47 a.m. on June 14, 2013), was among the first to test voice-banking in Minnesota and is now an outspoken proponent. Klinkhammer, 54, a gregarious former social worker from Albertville and bulky ex-college football player, has a bellowing laugh and a penchant for off-color jokes and puns. The technology enables him to keep speaking through an iPad, often late in the evening, when his throat muscles cramp and his voice starts to slur.
During a recent visit with his 28-year-old daughter Chelsea, Klinkhammer pulled out his iPad and began tapping out a few of his favorite puns as his daughter watched with amusement.
“A guy just threw milk at me. How dairy!” he wrote, throwing his head back in laughter. The words “I love you, Chelsea,” came out of a small speaker attached to his iPad with virtually the same inflection as his regular voice.
“That is definitely your voice, Dad,” gushed Chelsea, hugging him.
“This illness can come to define you because it takes away so many outward physical aspects of your being,” Klinkhammer said. “But voice banking epitomizes the idea that this illness can’t take everything away, because it can’t take away your voice.”