It’s a testament to the enduring power of The Beatles that their final new single, Now And Then, brought the world together yesterday in a moment of joy, love, hope and unity, 53 years after they split.
A global listening event elicited an outpouring of emotion from fans as the latest cutting-edge AI technology helped the band achieve the impossible and reunite one last time. Built around a haunting but sparse John Lennon piano demo from 1978, Now and Then plays as an emotional love letter between old friends echoing across the decades, its sentiment resonating with anyone who has ever lost a loved one.
Forty-five years after its composition, machine-assisted learning developed by Academy Award-winning director Peter Jackson has freed Lennon’s vocals from a mono cassette, converting it into startling studio quality. For anyone familiar with the original demo, hearing Lennon’s isolated vocals on the official ‘making of’ documentary this week was a genuine spine-tingling moment:
Paul McCartney, George Harrison and Ringo Starr first worked on the song in 1994 and 1995 when Lennon’s widow Yoko Ono gave them a batch of her husband’s unfinished, unreleased material to work on as part of the multimedia Beatles Anthology which spawned three CDs of Beatles rarities, outtakes and live recordings from the sixties and a major television documentary series.
Working with former ELO frontman Jeff Lynne as producer, the three Beatles added new vocals and instrumentation alongside Lennon’s vocals on ‘Free As A Bird’ and ‘Real Love’. Both became major top 10 hit singles around the world.
‘Now And Then’, however, presented significant problems. Not only was the song unfinished, but the cassette had an annoying electrical buzz running through it. It also proved impossible to turn up Lennon’s vocals without also raising the volume of his piano. They spent several hours on the track, recording new parts and completing a rough mix, but the challenges proved insurmountable.
With the Beatles exhausted after the emotion of the reunion, the tape went back on the shelf for nearly a quarter of a century, its completion seemingly out of the question after Harrison’s death in 2001.
McCartney, though, could never quite shake off the song. For him, it held a much deeper personal connection, calling back to mind the very last words Lennon said to him in person in the hallway of the Dakota building in New York: “Think about me every now and then, my old friend.” He remained quietly determined to bring it to completion but had no clear way to make that dream a reality.
Finally, the technology Jackson and his team developed during the landmark Disney+ Get Back series offered a solution.
Reviewing 60 hours of film footage and over 150 hours of audio capturing the Beatles talking, composing and recording throughout January 1969, Jackson’s WingNut Films developed machine-assisted learning (MAL) audio technology that allowed them to take any soundtrack and ‘demix’ or split all the different components into separate tracks, isolating individual instruments and vocals and all the individual voices within the Beatles’ conversations. One notable revelation was a candid conversation between Lennon and McCartney in the canteen at Twickenham Film Studios, previously obscured by the clatter of kitchen staff clearing china teacups and plates.
The technology was used again on the 2022 remix of The Beatles’ groundbreaking Revolver album and to allow McCartney to ‘duet’ with Lennon’s isolated vocals in an emotional performance of ‘I’ve Got A Feeling’ at Glastonbury in Summer 2022.
Behind the scenes, Jackson’s sound team, led by Emile de la Rey, were hard at work on ‘Now And Then’, applying the same techniques. Separating Lennon’s vocal from his piano, giving it its own track in the mix and preserving its clarity and integrity suddenly provided the key to finishing the song. It was as if Lennon was there, said Starr.
McCartney explained: “We could lift John’s voice without lifting the piano, which had always been one of the problems. Now we could mix it and make a proper record of it.”
The four-minute track opens with Lennon’s plaintive vocals, crystal clear, underpinned by McCartney matching his original piano part note-for-note.
Rumours that McCartney had used AI to ‘de-age’ his voice proved unfounded. Instead, his 83-year-old voice joins his old partner in harmony one last time, a Lennon frozen in time. The contrast adds real emotional heft.
With electric and acoustic guitar recorded by Harrison in 1995, the song builds with a new drum part from Starr and a fresh weaving bassline from McCartney, who also added a soaring slide guitar solo inspired by Harrison.
The two surviving Beatles also contributed backing vocals, with additional vocals from the original recordings of ‘Here, There And Everywhere’, ‘Eleanor Rigby’ and ‘Because’ woven into the song using the ‘mash-up’ techniques perfected during the making of the LOVE show and album.
Giles Martin, son of original Beatles producer George, provides the perfect orchestral arrangement, glueing the track together with overtones of ‘Eleanor Rigby’s’ stabbing, rhythmic strings and ‘I Am The Walrus” swooping, psychedelic colours, all while packing an emotional punch. Far from the melancholy original demo, the song is ultimately bittersweet yet beautifully uplifting.
Yet when news McCartney first teased the track in a BBC Radio 4 interview in June and hinted at the use of AI, it sparked fears he had used the technology to create a fake Lennon.
The backdrop was a recent flood of hundreds of simulated Beatles ‘songs’ appearing on YouTube, created using freely available AI software. Anyone who has ever imagined McCartney’s sixties voice ‘singing’ lead vocals on the Beach Boys’ classic ‘God Only Knows’ can now have their curiosity satiated…
Or ‘Lennon’ tackling an Oasis nineties anthem…
While impressive, for now, the results resemble an autotuned cover band performance. In the future, though, such fakery opens up serious ethical and moral questions around copyright and historical accuracy.
Talking about this, McCartney told the BBC’s Martha Kearney: “I’m not on the internet that much [but] people will say to me, ‘Oh, yeah, there’s a track where John’s singing one of my songs’, and it’s just AI, you know? It’s kind of scary but exciting because it’s the future. We’ll just have to see where that leads.”
Verity Harding, a globally recognised expert in AI, technology and public policy and lifelong Beatles fan, recently explained on the excellent I Am The Eggpod Beatles podcast why McCartney’s initial BBC interview had attracted so much attention.
“I think what’s happening is there’s just a big difference in different types of AI. So, AI is now an umbrella term, and it kind of covers a bunch of very different techniques and pieces of software. What the Beatles have been using is this kind of AI-enabled or machine learning-enabled ‘demixing’ technology. They teach the AI programme this is a piano and this is a voice, and this is a drum, and it can learn what those things are and separate them out, so you can then have each one individually.
“But people have confused that with a different type of AI which is called generative AI, and that’s when something completely new is generated.
“There’s two different ways it works. There’s voice cloning, where it takes your voice and turns it into a different voice, which is a kind of souped-up auto-tune. Then there’s a version where it kind of learns from lots of examples. And it is trained on that to be able to replicate it itself. So, you can teach it. This is what John Lennon’s voice sounds like. It can then create that in whatever way you want it to.
“So, it’s really amazing technology. But it’s obviously being used in some really exciting ways and some really concerning ways. One of the more concerning ways, I think, is when it’s being used to kind of crossover with somebody’s existing creative IP. And there’s a lot for the industry to think about when it comes to that.”
Ultimately, said Harding, the use of AI in music is not something to panic about but a natural progression. She quoted from a recent interview about AI with David Guetta: “He thinks it’s just a tool. He’s not worried about it. He said it can’t replace taste. It’s just another tool people like him can use, and it will reduce music production cost, but he very much was kind of describing it as just another tool in his toolbox.”
She understands people’s concerns when it comes to AI and the Beatles above any other artist, in part because of their huge cultural significance but also because of how much they mean to people.
But she concluded: “My prediction is, human beings are always going to love what other human beings do. The Beatles are just an astonishing, genius talent, but they are also very, very human, and their humanity is part of what we love about them.”
Some point to ‘Now And Then’s’ long gestation, its sessions separated by history, and its reliance on studio ‘trickery’ to assert it cannot be a real Beatles recording.
But McCartney is insistent Lennon would have readily embraced modern recording techniques. As Harding notes, the Beatles “were at the forefront of using and abusing technology”, among the very first to use the studio as an instrument. EMI’s Ken Townsend created Artificial Double Tracking for them to avoid having to continually re-record vocals to create multi-layered vocal effects. For ‘Tomorrow Never Knows’, Lennon’s vocals were put through a Leslie speaker cabinet and double-tracked, while the track was adorned with tape loops. ‘Strawberry Fields Forever’, meanwhile, merges two separate takes recorded in completely different keys and tempos.
‘Now And Then’ was a song born out of love, but one carrying an enormous weight of expectation. Yet McCartney and Starr, inspired by one last burst of joint creativity with their former comrades, have pulled off something befitting the band’s enormous legacy to provide a suitable closing chapter.
Said McCartney: “There it was, John’s voice, crystal clear. It’s quite emotional. And we all play on it, it’s a genuine Beatles recording. In 2023 to still be working on Beatles music, and about to release a new song the public haven’t heard, I think it’s an exciting thing.”