With A Little Help From AI, The Beatles Are New Again
Image Source: BBC
The Beatles were and still are a huge cultural phenomenon. Though they primarily only existed as a group from 1960-1970, their music continues to inspire and entertain audiences and musicians alike. When they broke up in 1970, there was a huge musical void left that others have attempted to fill. The ex-Beatles were always asked if and when the group would reunite. Though there were partial reunions throughout the 70s, there was never any true Beatles reunion. When John was murdered in 1980, that seemed to dash all hopes of a reunion. However, in 1995, a miracle happened. The then-surviving Beatles, Paul McCartney, George Harrison, and Ringo Starr, reunited to work on The Beatles Anthology project, and part of that project was the release of two new Beatles songs produced from demos recorded by John Lennon: “Free As a Bird” and “Real Love.” A third song, “Now and Then,” was meant to be released as well, but there were too many technical problems with it, and George vetoed it, so it was shelved.
Fast forward nearly 30 years later, and another miracle has occurred. Paul McCartney has recently announced that a new “final” Beatles song will be released later this year. Though no real details have been announced at this time, McCartney did mention that the track used artificial intelligence in its creation. With the advent of AI creation tools such as Respeacher and ChatGPT, it was immediately assumed that this song would use this technology to create artificial versions of John Lennon and the late George Harrison (who passed away on November 29th, 2001). However, this is not the case at all, and this same technology has been used before on the Peter Jackson-directed Get Back miniseries and the recent Revolver re-release: de-mixing.
RELATED:
Image Source: Best Classic Bands
This de-mixing technique was used extensively in The Beatles: Get Back miniseries. Peter Jackson and his team developed a machine learning based technique to extract the band’s voices from the mostly mono tapes used during the production of the original Let it Be documentary. The group would intentionally turn their amps up during a conversation so Lindsey-Hogg and his team couldn’t hear what they were saying. The computer learns what a guitar, bass, drum, and voice sounds like and uses that information to clean up the tracks and extract only the audio needed. This technique was also used to allow McCartney to “play” with his bandmate John Lennon again on the song “I’ve Got a Feeling” during his 2022 Got Back tour (pictured above). Giles Martin also used this technology to remix the 1966 Revolver album, which was recorded using mono 4-track recording technology, making it seemingly impossible to remix in the way albums such as Sgt. Pepper’s Lonely Hearts Club Band, The Beatles (also known as The White Album), and Abbey Road were remixed, as they used more advanced techniques and were recorded in such a way that the instruments were more separated. Now, this same technology is being used to reunite The Beatles.
Paul McCartney initially announced this during an interview on BBC’s Radio 4. In it, he was primarily promoting his new book, 1964: In The Eyes of the Storm. During this interview, he asked about AI-generated covers of his songs that would feature, say, a younger version of his voice or John’s voice singing one of his songs. McCartney said that technology was “kind of exciting” and “scary” but said that musicians might have to learn more and more to work with AI as it seems to be part of the future. It was during this statement that he mentioned that AI was used to help make this new Beatles record, leading some to speculate that John’s voice would be artificially generated. McCartney later clarified on Twitter that this was not the case and that the technology was being used to clean up the demo and extract his voice and piano, making them separate tracks so that the song could be mixed as normal. Below is his full statement, which confirms involvement from both George Harrison and Ringo Starr:
Image Source: GQ
Ringo Starr also weighed in on the matter. He says that The Beatles would never fake Lennon's vocals on the track and that AI was simply used to clean up previously recorded material. Starr also confirmed that the new track would feature George Harrison's vocals as well, stating:
“This was beautiful…it’s the final track you’ll ever hear with the four lads. And that’s a fact.”
Even John Lennon's son with Yoko Ono, Sean Ono Lennon, stepped into the ring to clarify how AI was being used. When a fan accused McCartney of creating an AI voice model of his father's voice, he replied to clarify that that was not the case:
The use of AI within media is a bit more nuanced than it may seem. While the focus right now is rightly on how it can be used to put words unto people's mouths or how work is being stolen to create new works, there is a lot of good AI. can do as well, such as clean up audio, and that is precisely what is happening here. There's no voice cloning of Lennon and Harrison being done. What you will hear on the track is perhaps something you will never hear again: the authentic performances of John, Paul, George, and Ringo together one last time.
READ NEXT:
Sources: Twitter, [1], [2], CNN, Best Classic Bands, YouTube, Open Culture, Noise 11
Latest News
Featured
