BATON ROUGE, LA (WAFB) - The sun is new in the sky and Donna Britt, still not a morning person, is out for a recording session at LSU's Speech Language and Hearing Clinic in Hatcher Hall.
Students are beginning to stream onto the sidewalks as they head to their early morning classes. These are one hour sessions to lay down what some have called Donna's "major professional signature," her voice. About 30 years ago, a New York talent agent told Donna she had one of the best female TV voices in the business, but that TV voice is also her personal one. Britt says she knows with ALS, she will eventually lose it when the muscles in her throat start dying.
So in a tiny, sparse recording booth, LSU grad student, Emily Watkins, fires up the software and Donna records phrases and sentences that are somewhat peculiar. Many are sentences from classic books like Little Women, Wizard of Oz, and more.
"I need to acquire a brain," said the Scarecrow.
"We try to get up to 1,600 sentences or phrases, if you will, that what we use is ModelTalker. So we can work with a minimum amount, I believe it's 400, if that's all we can provide, but the other option we can provide, is that a family member who has a similar type voice pattern as you, they can use," said Sara Mele, an LSU Speech & Pathology instructor supervising students working in the voice bank.
Donna says her goal is to take communication beyond the robot sound of the world's best known ALS patient, physicist, Stephen Hawking. His struggle to communicate literally inspired the rapid development of the talking machine. As Hawking developed a new handicap, scientists were eager to answer it and enable Hawking to keep lecturing and writing books. But Hawking's voice was destroyed before it could be recorded, and so his device uses a machine-derived voice.
Pursuing Donna's own synthetic voice, Watkins watches all the gauges labeled with pitch and intonation for instance. She tries to get the clearest recording possible.
On headphones, Donna hears a man enunciate and intone the phrase the way Donna should, the Watkins clicks record and Donna speaks, "Therefore, we must stop where we are." Donna muffs the final word and they both laugh. "Oops, OK," Donna says. She then repeats the phrase, "Therefore, we must stop where we are." These are clearly elocuted sentences that contain break apart sounds for building new words.
What do LSU students get out of this? Dr. Yunjung Kim, Associate Professor of Speech & Hearing Services answers, "Students can learn about speech analysis, speech acoustics, speech movement, and more importantly, how neurological conditions such as ALS can impact a communication ability."
Donna says she took to the web for samples of the finished product, and felt the ones she saw on YouTube were not that impressive. The synthetic voice still sounded like a computer Donna felt, just using their voice and a default setting on ModelTalker's software.
But Britt says, "My voice will be recorded while it's still here and as technology improves, which take months or years, I hope to hear it come out of a machine sounding more like the natural me."