It seems like with the advances in AI technology that in the near future old bootleg recordings such as any 1972 show will be able to be converted to a quality sound equivalent to a soundboard recording. This would not be altering the show but enhance the actual performance to a high quality equal to a genuine release. Is this a realistic thought?

Quote
yearsinthemaking
It seems like with the advances in AI technology that in the near future old bootleg recordings such as any 1972 show will be able to be converted to a quality sound equivalent to a soundboard recording. This would not be altering the show but enhance the actual performance to a high quality equal to a genuine release. Is this a realistic thought?

It would be. If it wouldn't be altering you wouldn't need AI. Such an AI "improvements" would be like guessing how it most likely had sounded back then. I'd rather stay with the old audience recordings than hear a computer generated concert in excellent quality.

But technically this is completely possible and the more effort you would put in it the more realistic sounding the result would be. You could separate the recording into different tracks and improve these individual Tracks. But every missing information from the original tape can only be recreated on a "best guess". I hope the Stones will never support such projects.



Edited 1 time(s). Last edit at 2026-02-01 00:39 by hockenheim95.

That's a fair question.

What Peter Jackson did for GET BACK in cleaning up the songs was/is a particular kind of AI, although supposedly not actually AI it's "deep machine learning" when it was being done, from what I understand, called Demixing. It was used on the mono recordings during rehearsals to separate the talking from the guitars.

[www.mixonline.com]

Giles Martin used demixing as well with/for the recent anthology 4 release with the new song, I don't recall if the other two singles were revamped with the same technology to clean them up.

[www.musicradar.com]

I suppose at some point people not in the industry will get ahold of the software somehow and be able to - like how people wish LADIES AND GENTLEMEN could have a better mix - clean up a muddy sounding album. Someone will be able to do it.

Imagine if Elvis' early catalogue was demixed. And a load of other artists from the mono to 2 track recordings era. It could be a new era for the music industry. As Giles Martin talked about not changing anything, imagine something by Elvis that's now multi-tracked all original recording as well as some kind of remix, like a club mix etc. Enhanced, altered, added to, whatever.

That would be amazing.

Quote
hockenheim95

Quote
yearsinthemaking
It seems like with the advances in AI technology that in the near future old bootleg recordings such as any 1972 show will be able to be converted to a quality sound equivalent to a soundboard recording. This would not be altering the show but enhance the actual performance to a high quality equal to a genuine release. Is this a realistic thought?

It would be. If it wouldn't be altering you wouldn't need AI. Such an AI "improvements" would be like guessing how it most likely had sounded back then. I'd rather stay with the old audience recordings than hear a computer generated concert in excellent quality.

Broadly I agree with hockenheim95, but....

The sole exception I'll make is if Peter Jackson's WingNut people are involved. The job they did on The Beatles recordings for the "Get Back" movie was simply unbelievable... plus it was respectful at the same time.

Importantly they took the original (rubbish) recordings and produced tracks for each instrument - and these tracks were then re-mixed by (usually) Giles Martin (ie a professional producer and his team). ie a 2 stage process - extract and improve the tracks, and then re-mix them.

Whether or not WingNut uses AI does rather depend on your definition. Generic AI seems to be trained indiscriminately on anything the developers can get their hands on. WingNut (reportedly) worked only with recordings of musical instruments and the specific requirement of perceived sound. ie Not Generic AI training. (There's a good case for saying that WingNut's developments are NOT AI - but that's another discussion.)

For the record, Peter Jackson now owns the originals of The Beatles Star Club recordings - which have to be considered classic bootleg stuff - and he is working on those. With the full knowledge of Apple. But no information about their release to the rest of us though!

(Edit: LOL! My posting took longer for me to write than gaslightstreet's! Great minds....)

Captain Corella



Edited 1 time(s). Last edit at 2026-02-01 00:52 by CaptainCorella.

Is it remixing though? Taking a mono recording and creating multi-tracks is technically having the ability to create a mix, and an original one at that, since mono recordings and, arguably, two-track recordings, aren't a mix - it can't be remixed. Sure, mic placement dictated who was heard the most, in mono, as well as how instruments were bussed to the two track was a mix in itself but permanent.

Four track recording, like the early Beatles and Stones etc have more instrumentation due to the bouncing of three tracks to one, two tracks to a second track and various punch ins on the remaining two free tracks, which was still done, apparently, with 16 track recording if one is to believe the track notes that are in BLACK AND BLUE.

Look at Fool To Cry - track 7 had Keith and Wayne overdub guitar parts yet tracks 15 and 16 had nothing recorded on them. Crazy Mama - track 9 has Billy, Keith and Ronnie on it yet track 13 is blank.

Imagine having the deep machine learning to demix TSMR by uploading into Pro Tools, do a however many tracks per song mix, with possibly editing some of the silliness down, with a 'as it was' mix and a padded/enhanced mix.

Not to make it sound modern, which is an extremely wide horizon now compared to even 1997, but to make it sound better.

Someone that can't stand the mix of EXILE could clean it up... fix the recent live album/live video mixes - it's wide open with possibilities. Unthought of 30 years ago is now possible. Not much different than various things recorded to VHS, like I did with certain sports moments, highlights, yet alone 1989 Atlantic City, which I crudely recorded to cassette so I could listen to it wherever, that are most likely now on YouTube and available on CD and AppleMusic etc, like Atlantic City.

As much as AI is annoying, and wrong, as in full of errors, utilising it for making music sound better - or have a sound at all ala GET BACK via Park Road Post Production in New Zealand - is a fantastic application of AI - or deep machine learning.

Forget SUNO. Demix your favorite Stones songs, remove Mick or Keith's lead vocal and record your own...

Another example is the amazing film, also by Peter Jackson, "They Shall Not Grow Old". Which took WWI film and digitally cleaned everything up. A purest might not approve of such tinkering but the results are fantastic.

I don't think anything "fake" would be created in such a recording, only enhancing what is there and poorly recorded on old equipment in the middle of a crowd

Quote
yearsinthemaking
Another example is the amazing film, also by Peter Jackson, "They Shall Not Grow Old". Which took WWI film and digitally cleaned everything up. A purest might not approve of such tinkering but the results are fantastic.

I don't think anything "fake" would be created in such a recording, only enhancing what is there and poorly recorded on old equipment in the middle of a crowd

With the original intact why would anyone get upset about improving something? The 2025 BLACK AND BLUE sounds better than the 1976 original.

Creating a multi-track for mono audio is incredible.

We have also lots of the old tracks that were cleaned up and remixed in stereo on Youtube. But then, artificial stereo sound - back then only a slight improvement - was common on the official Rolling Stones albums since the 1970s. That means that they themselves had the original mono tracks altered.

Just for comparison:

Obviously there is already a seies of AI created 'lost soul albums from the 60's' on Youtube, for example

[www.youtube.com]

If you listen to the audience recordings of the legendary 1973 Tour of Europe on Youtube you will notice that many of them were cleaned up and sound much better than on the original bootlegs.



Edited 1 time(s). Last edit at 2026-02-03 22:38 by saltoftheearth.

Ladies and Gentlemen is an official release and the sound is bad.AI might clean it up.The guitars particularly.

Quote
GasLightStreet
That's a fair question.

What Peter Jackson did for GET BACK in cleaning up the songs was/is a particular kind of AI, although supposedly not actually AI it's "deep machine learning" when it was being done, from what I understand, called Demixing. It was used on the mono recordings during rehearsals to separate the talking from the guitars.

[www.mixonline.com]

Giles Martin used demixing as well with/for the recent anthology 4 release with the new song, I don't recall if the other two singles were revamped with the same technology to clean them up.

[www.musicradar.com]

I suppose at some point people not in the industry will get ahold of the software somehow and be able to - like how people wish LADIES AND GENTLEMEN could have a better mix - clean up a muddy sounding album. Someone will be able to do it.

Imagine if Elvis' early catalogue was demixed. And a load of other artists from the mono to 2 track recordings era. It could be a new era for the music industry. As Giles Martin talked about not changing anything, imagine something by Elvis that's now multi-tracked all original recording as well as some kind of remix, like a club mix etc. Enhanced, altered, added to, whatever.

That would be amazing.

Your post was really helpful in understanding the AI processes.

Thanks for the information!

Duck

“AI Bootleg recordings”, I myself am not interested in the so called improving sound of bootlegs by AI cause I stick with the original and realistic recordings even the bad sounding, the screaming and the madness wich are more real compared to the more fake AI! I love my bootlegs

Here is an interesting audio snippet of the Beatles that had an AI makeover. It seems like nothing was altered. Same performance only enhancement of the audio. Thoughts???

[youtu.be]

Couldnt AI clean up the snow from the footage ??????



ROCKMAN

I believe Mayflower did it with Stones In The Park. It sounds better than it ever did to me.

Quote
yearsinthemaking
Here is an interesting audio snippet of the Beatles that had an AI makeover. It seems like nothing was altered. Same performance only enhancement of the audio. Thoughts???

[youtu.be]

It just sounds like a tweaked EQ.

Quote
runaway
“AI Bootleg recordings”, I myself am not interested in the so called improving sound of bootlegs by AI cause I stick with the original and realistic recordings even the bad sounding, the screaming and the madness wich are more real compared to the more fake AI! I love my bootlegs

agree %100 !!!!!!

I've heard some amateur experiments attempting to clean up audience recordings. Mixed results so far, to be fair - sometimes it actually felt like an improvement, but some other quieter passages were totally off - sometimes it reminded me of the awful 80's Swinging Pig releases with their "noise-reduction" mastering. Win some, lose some, I suppose.

While I'd still prefer an audience recording untampered, I reckon in a near future AI could be a very interesting tool. Let's look at it like Photoshop - you could have atrocious results if you're not skilled at it and haven't developed a well-balanced taste, but it could also be an undeniable tool if you know what you're doing.

AI is just Technology. You may ask AI to replace guitar bits by a person with bits from other songs, you may change pictures, videos, sound, text, it is perfect for fraud, and as I said in a separate thread, IORR is not a technology site:

IORR posting policy related to spam, chatbots, ChatGPT and AI - UPDATE JAN 20, 2026 (IORR)

This discussion is about bootlegs, so it is moved to the Hot Stuff section.

Bjornulf

The Rolling Stones are famous for playing fantastic live performances. To attend their shows, to experience their spirit is, what's it's about. Their shows have been recorded by fans
and pros. Bootlegs, audience, sounboard or IEM recordings are nothing more than trying to catch that spirit.
No AI technology is going to be able to replace this.
A person, who never saw them life may not be able to get the point. But we all here at this forum have seen them so many times.
For me one of their secrets is their ability to be in the very moment when they perform and to celebrate this with the whole band and the audience.
They share the joy of being alive.
I can't imagine AI could catch any point of this experience.



Edited 1 time(s). Last edit at 2026-02-04 23:31 by micha063.

I'd say it all depends on the source of the recording, which if its a real bad aud or bad soundboard then AI software can only do so much. Inserting fake music parts would be a whole different thing.

Overall IMO there's some great tools out there to tweak things like bringing up or down the vocals, drums, etc.

I'm definetily against any kind of faking or altering anything, but if let's say AI could reduce the screaming and bring the music up front on an audience recording I think that's good! ... I yet have to hear some convincingly enough good result, though.

But the way I see it is like matrix mixes of two differente sources or IEM spliced with audience recordings. Ideally, I'd rather a straight away "clean" recording of any kind, but if IA or whatever EQ trick could clean anything up, that would be welcome!

Just what we all need. AI remasterbaters. No......

Quote
Taylor1
Ladies and Gentlemen is an official release and the sound is bad.AI might clean it up.The guitars particularly.

The desire to improve poor AUD recordings, like those from 1972, to near-perfect SBD quality is still a very distant dream. Theoretically, it's possible to combine the "information" provided by an AUD with the sound quality of another SBD from
1972 and create an artificial SBD. This is very theoretical, but I could imagine it being possible. However, there's currently no technology that could implement this.

What Jackson does is something completely different: taking existing audio and dividing it into different channels in order to then remove or enhance certain
channels. This can also be called AI, but in principle, nothing is artificially created; rather, AI is used to deconstruct existing recordings.



Edited 1 time(s). Last edit at 2026-02-11 14:11 by falo01.

It will definately be possible in the near future.
I dont have a problem with it.

Personally I don't mind it as long as the original band members are in the recordings. They can even have Keith or Ronnie playing missing bass parts for example. Or if one them can play drums why not.

Quote
frankotero
Personally I don't mind it as long as the original band members are in the recordings. They can even have Keith or Ronnie playing missing bass parts for example. Or if one them can play drums why not.

How will Keith "playing drums" when he doesn't play drums be anything but AI garbage. Sounds like something Rskinno would do after he reads a help screen on some AI software program.

...I don't want to hear Ringo drumming on Satisfaction...