Posted by Bud on February 26, 2010 at 15:38:01:
Tech firm gives Roger Ebert his own voice
February 26, 2010 12:57 PM
After his second surgery, when he learned he would no longer be able to speak, Roger Ebert started writing notes.
But the iconic film critic soon found this took too long to keep up with the normal flow of conversation.
"There is a point when a zinger is perfectly timed, and a point when it is pointless," Ebert wrote in an August 2009 blog entry.
Last summer, Ebert found the Web site of CereProc, a company based in Edinburgh, Scotland that uses "text to speech" technology to provide people with voices that "sound real" and "have character," according to the company's Web site.
Using mostly audio from Ebert's own voice on DVD commentaries of "Citizen Kane" and "Casablanca," the company created a voice that sounds like the one he lost.
On Friday, Ebert demonstrated his voice during a taping of Oprah Winfrey's show, sounding like himself for the first time in nearly four years since his battle and recovery from thyroid cancer left him voiceless, according to a press release from the show's producer.
The show, which will air Tuesday, includes Ebert giving the talk show host his 2010 Oscar picks and allowing her cameras to follow him for a day.
A profile of Ebert in the March issue of Esquire detailed how losing his voice has amplified his writing. But on his blog, Ebert wrote about how he still missed his own voice. He described a synthetic voice software for his Mac as "a godsend."
"Most of the voices, however, left a lot to be desired," he wrote.
Ebert's new voice is still being fine-tuned, and the project will not be completely finished for a few months, said Matthew Aylett, chief technical officer for CereProc.
And while the new voice sounds like his old one, it occasionally makes errors, Aylett said. It has difficulty pronouncing unusual proper names and sometimes the intonation does not sound natural, he said.
"It sounds like him," he said. "But it will sound better as we add more audio information to it. The more data we have, the smoother and the more accurate the voice will become."
To build a voice, the company needs about 300,000 different sounds from the speaker, Aylett said. The company needed between three and five hours of audio from Ebert speaking to produce the voice.
"Roger has many years of experience in broadcasting," Aylett said. "Obviously we couldn't record him but he did have a lot of audio material we could use to build his voice."
Aylett said this is the first time the company has produced a synthetic voice that sounds like the old voice of the person using it.
"Your voice is very much a part of who you are, your history and identity," he said. "We're giving Roger back his power of speech, to a certain extent."
Last summer, Ebert said he had big plans after discovering CereProc.
"I can see my own voice hosting online or telecast video essays," Ebert wrote. "I am greatly cheered."
-- Gerry Smith