Rupal Patel, a speech scientist behind the Human Voicebank Initiative, wants to diversify and liven up the range of computer-generated voices available to people physically unable to speak. Megan Garber explains the process:
It works like this: Volunteers come to a studio and read through several thousand sample sentences (sourced from books like White Fang and The Wonderful Wizard of Oz). Patel, [collaborator Tim] Bunnell, and their team then take recordings of a recipient’s own voice, if possible, to get a sense of its pitch and tone. (If the recipient has no voice at all, they select for thing like gender, age, and regional origin.) Then, the team strips down the voice recordings into micro-units of speech (with, for example, a single vowel consisting of several of those units). Then, using software they created—VocaliD, it’s called—they blend the two voice samples together to create a new, lab-engineered lexicon: an acoustic collection of words that are at the disposal of a person who needs them to communicate.
Randy Rieland elaborates:
The project’s website, VocaliD.org, has both a sign-up page for donors and another for those hoping to get a personal voice. The latter must submit their names and other relevant information such as their speech ability, which can range from “completely non-vocal” to “can make sounds but not words” to “can use some words for communication.”
While only a handful of voices have actually been created during the project’s infancy, more than 10,000 people already have volunteered to be voice donors, Patel says. “Several hundred” others, she says, have signed up to get new voices. … Her vision is to collect a million different voice samples by 2020. But already her work is making an impact. The site features an audio file, only two sentences long, provided by a young woman described as having a “severe speech impairment.” Her words are as clear as day:
“This voice is only for me. I can’t wait to try it with my friends.”
The Dish 