Beatles legend Sir Paul McCartney shared that “the final Beatles record” is set to be released with an AI-isolated vocal from the late John Lennon.
Speaking to BBC Radio 4’s Today program in the UK, McCartney shared that AI had assisted in isolating John Lennon’s voice from an old demo track, allowing him to resurrect Lennon’s original vocal for a duet. “We just finished it up and it’ll be released this year,” he said.
McCartney didn’t disclose the song’s title, but it’s presumed to be John Lennon’s 1978 piece, “Now And Then,” which was intended to be a ‘reunion song’ for the Beatles’ autobiographical Anthology TV series in 1995.
The demo of the song was a gift from Lennon’s widow, Yoko Ono, to McCartney. It was part of several songs on a cassette labeled “For Paul,” produced by Lennon just before he died in 1980.
Two of these tracks, “Free As A Bird” and “Real Love,” were enhanced by producer Jeff Lynne and released in 1995 and 1996, marking the Beatles’ first “new” material in 25 years.
The Beatles attempted to record “Now And Then” but later abandoned the project. McCartney told BBC Radio 4 that George Harrison disapproved of the song due to the unsatisfactory sound quality of Lennon’s vocals, “George didn’t like it. The Beatles being a democracy, we didn’t do it.”
McCartney has often admitted he wished to complete the song, including in a documentary on Jeff Lynne in 2012, where he stated, “That one’s still lingering around. So I’m going to nick in with Jeff and do it. Finish it, one of these days.”
How AI resurrected Lennon’s voice
During the production of Peter Jackson’s “Get Back” Beatles documentary, dialogue editor Emile de la Rey used AI to isolate the band’s voices from background noises and even their own instruments, creating clean audio from old recordings.
With this technique, McCartney was able to “duet” with Lennon on his recent tour, and it helped create new surround sound mixes for the Beatles’ Revolver album last year.
McCartney described the process, “He [Jackson] was able to extricate John’s voice from a ropey little bit of cassette…We had John’s voice and a piano and he could separate them with AI. They tell the machine, ‘That’s the voice. This is a guitar. Lose the guitar’.”
He added, “It’s kind of scary but exciting, because it’s the future. We’ll just have to see where that leads.”
AI’s role in music is expanding. It wasn’t long ago that London-based artist “patten” released the first 100% AI-generated album. In early May, fraudsters sold AI-generated deep fake music on an online forum.
This isn’t an example of ‘fake’ or ‘AI-generated’ audio. Here, AI extracted a real voice from a recording and cleaned it up by manipulating sound wave data. Such techniques have been used in audio restoration for decades, but modern AIs produce considerably cleaner results.