LONDON -- Yes, Elvis lives. And soon he may be singing your song.
Ken Lomax of the University of Cambridge has developed a way to reproduce the singing and speaking styles of performers, living and dead. So far, Lomax has generated voice "templates" for opera stars Maria Callas and Kiri Te Kanawa -- and the king of rock 'n' roll.
Lomax's "voice morpher" builds a template of a performer's voice from recordings. It captures features of the performer's distinctive singing style, such as how the words of a song are pronounced, the tone of the voice and the characteristic timing of certain phrases.
Lomax is working on applications for his technology, such as a karaoke machine that can make a person sound like his or her favorite singer, and a synthesized voice for computers that can mimic film stars.
He envisions a karaoke machine using the morpher to convert the voice of anyone at the mike into that of Elvis -- or any other performer. The machine would preserve the timing of the karaoke singer but change the pitch and tone to make the person sound like the King.
A spokesman for Isis Innovations, which is helping Lomax find commercial partners, says the system has two kinks to work out -- the time it takes to generate the template and the processing power required to blend the actual and electronic voices in real time.
Still, Sony, Sharp and IBM are evaluating Lomax's system with a view to commercial products.
Pub Date: 3/30/98