Elvis Presley and John Lennon are set for a spooky comeback. A research project at Oxford University is developing the perfect electronic replication of their voices. PhD student Ken Lomax, at St Hugh's College, is working on a singing synthesiser which might one day allow diehard fans to write personalised songs and have their dead heroes sing them.
Mr Lomax, whose project is called Singing Analysis and Synthesis, is working with local choristers but says that it should be possible to analyse, and so replicate, the recorded voices of greats such as Elvis, Lennon and Buddy Holly.
He has hit a serious snag in record companies' reluctance to release the vocal-only tracks he needs to "lift" voice patterns free of background music.
"It throws up all sorts of possibilities such as having Elvis and Lennon sing a duet," he said. "It also raises legal conundrums, such as people having to patent their voices and even the prospect of making someone say something they didn't."
A programme installed on a powerful computer scans voices and breaks them down into their constituent parts. The programme then reassembles these characteristics to create a voice replication which not only fools the human ear but, ultimately, passes any existing electronic voice analysis.
Mr Lomax said: "There's no reason why such a programme couldn't work on a PC - such synthesisers may one day be available in high street shops. It could add a whole new twist to home entertainment."