Speech Technology Magazine

 

VocaliD Launches BeSpoke Voices for Assistive Devices

Bespoke blends vocal DNA from those who can't speak with recordings from those who can to create unique voices.
Posted Aug 8, 2016
Page1 of 1
Bookmark and Share

VocaliD, a speech technology company that creates personalized digital voices, today launched its online BeSpokevoice studio.

Supported by the National Science Foundation and the National Institutes of Health, VocaliD's innovation is based on a discovery by founder and CEO Rupal Pateland her colleagues that even a single vowel contains enough vocal DNA to seed the voice personalization process.

The BeSpoke process captures the unique vocal identity and blends it with recordings from a healthy speaker, matched by gender, age, and accent, within the company's Voicebank of more than 14,000 contributors. This blended voice is then used in the recipient's electronic communications device.

"The inspiration for VocaliD came when I saw a young girl and older man communicating using their assistive devices but with the exact same adult male voice," said Patel in a statement. "We wouldn't dream of fitting a little girl with the prosthetic limb of a grown man, so why then the same prosthetic voice?"

More than 10 million globally are unable to speak and rely on assistive devices.

For David Manzo, president of the Lexington, Mass.-based Cotting School for students with special needs, VocaliD is "breaking down barriers for children with complex challenges to communication."

"I'm observing first-hand the communication and self-esteem benefits of VocaliD's technological breakthrough on our students," Manzo said in a statement.

Sridhar Tayur, Ford Distinguished Research Professor at Carnegie Mellon University, who serves on VocaliD's advisory board, said, "the discovery of a person's vocal DNA combined with a crowdsourced Human Voicebank and development of 21st century hyper-scalable digital synthesis algorithms is a trifecta."


Page1 of 1