22/01/2013 - CereProc's Wonder Emporium of Text-to-Speech - Cloning Voices
Anytime anybody impersonates you, it's a great compliment.
Robert Wagner (Film Actor)
Our voices are part of our identity. Mimicking voices has long been an element of satire and comedy. On a slightly darker note, the duplication of voices has also been been used within science fiction to make us feel uncomfortable. From Invasion of the Body Snatchers to Terminator II, the theme of a robot or alien stealing a person's identity - and thus adopting their voice - is a common one.
While voice cloning is used to entertain - whether through humour or fear - being able to copy a person's voice offers a whole host of positive uses in today’s society. Many people lose their ability to speak for medical reasons, and many would love to be able to use their original voice to connect with their world through computer-aided communication.
Stephen Hawking is a famous user of speech synthesis. But the voice he uses is not his original voice. It is an 80s speech synthesis system which is now indelibly linked to him as a world renowned figure. If we had access to audio recordings of Stephen Hawking speaking before he lost his voice, we could now duplicate his real voice.
This is exactly what CereProc did for the famous movie critic Roger Ebert, who has discussed what it means to lose your voice on Oprah and other programs. Here's a clip from CBS News:
To test your speech recognition abilities, we've duplicated some famous - and some not so famous - voices. Can you tell us who they are?
Last year’s 2012 Coachella Valley Music and Arts Festival in California saw a 3D holographic Tupac Shakur perform with Snoop Dogg and Dr. Dre, and the response from fans, music industry types and technology industry experts alike demonstrated the growing interest in this interpretation of cloning. Keeping this in mind, and focusing on our role as developers of speech technology, we're very excited that by copying the voice we can then collaborate with peers from other sectors to create complete virtual versions of a person, complete with personality.
For a bit of fun have a look at the Bush-o-Matic made with the help of Idyacy Ltd:
http://www.idyacy.com/cgi-bin/bushomatic.cgi
Our next exhibit - Don't get mad get snippy! - explores the fundamental issue of giving synthetic voices emotions, another key part of building a synthetic voice with personality and relatable identity. Don't miss it!
In case you missed them, here's the Wonder Emporium Part 1 and Part 2.