Spencerville Student Helps Others Find Their Voice

Story by Heidi Whetmore

As Spencerville Adventist Academy (SAA) sophomore Autumn Uhrig watched a YouTube video one morning, something piqued her interest. Millions of people who can’t speak rely on text-to-speech devices to communicate. Most devices have the same vocal sound for all users, regardless of age or gender. Uhrig learned of VocaliD, which, according to their website, is “the human voicebank of the world.”

VocaliD records human voices, mixes them with sounds a non-speaking person can make and blends them to make a unique humanized voice in which the machine talks for them. Uhrig wanted to do her part in helping someone find their voice. She visited the website, recorded a couple of sentences and waited to see if there was anyone her age who needed assistance. Months later she received an email that a young girl, Tesa, was a good match for her voice. Uhrig now needed to record 3,488 sentences to get the job done.

As expected, this process took hours. “I would come home from school every day, do my homework, and then go to a quiet room to record as many sentences as I could before bed. At the end of each session, my throat would hurt a little bit, but it was worth it,” says Uhrig. “I recorded the sentences as quickly as possible because [VocaliD] hoped to give this little girl her new voice by Christmas!”

Uhrig finished the sentences in time. Tesa got her very own voice, and other girls will benefit as well. Uhrig’s recordings can be matched with anyone who thinks her voice is compatible with theirs. “Currently there are three young women considering my voice as a match,” says Uhrig. “Though it took a long time to do—it was actually pretty exhausting at times—it was well worth it. Those women will get their own voice, and I had a part in giving it to them.”