Voice synthesis has really come a long way. For the longest time, Stephen Hawking's robotic voice has been the standard. But this week, movie critic Roger Ebert, who lost his jaw and voice to cancer, appeared on the Oprah Winfrey show, demonstrating his new synthesized voice, which sounds remarkably like him. The voice was made possible by the Scottish company CereProc, located in Edinborough.
This interview on NPR with Matthew Aylett, CereProc's chief technical officer, gives the basics of how the company makes synthesized voices. I recommend listening to the interview, instead of reading the transcript so you can hear the synthesized voices, including one that sounds like George Bush.
Aylett says that creating the voice is fairly straightfoward:
"You take a lot of audio from a speaker. You then cut that up into tiny little pieces. Each piece is a little sound. So for example, cat would be made up of three sounds, /k/, /a/, and /t/. In order to then produce a new sentence, you then take those sounds, you rearrange them, and you stick them back together again."
Below is a video clip from the Oprah show. It's almost seven minutes long. I recommend watching the whole thing so that you can hear Ebert's current synthesized voice, which is fairly generic, and then hear the new synthesized voice, which really does sound like Ebert.
It makes me wonder if Stephen Hawking would consider getting a voice synthesizer that sounded less robotic. And yet, he's had it so long, that it has become distinctly HIS voice.
AP Photo/Evan Agostini
Tags: Biotechnology
comments ( )