Militant said:

lurker316 said:

Yesterday I listened to the Peach and Black podcast's interview of Niko Bolas (the engineer from the Orgiinals). He suggested the oldest tape is actually not in the most danger because its chemicat composition was a better qualtiy. It's the slightly less old stuff recorded on tape with a different, inferior composition that's in greater danger.

That actually makes a lot of sense when you consider that the first two albums were done in pro studios.

Prince blew his three album budget on mostly the first album and then the second. That's why Dirty Mind sonically sounds muddy - but given the style of the songs, it worked. It was recorded in a low budget leaky basement in his house.

So it's likely the stuff recorded at his house, between I guess 1980 and around 82 or so when he mostly started working at Sunset Sound (and before 1999 became a hit and he had more money to work with) where lesser quality recording material was used.

I suspect the most obvious example we have so far of this, is the version of "Wouldn't You Love To Love Me?" that ended up on Originals. There was only so much they could do with what they had, and thus it's sonically worse than the rest of the album.

.
WYLTLM sounds bad because it was likely sourced from compact cassette and it's drenched in reverb (it almost sounds like it was put through a spring reverb, lol), not because it was tracked on a lesser quality 2" tape. Howe & Co. probably either couldn't find the multitrack (maybe it's not marked on its box) or it was badly damaged and they couldn't save it.
.
Generally speaking though: hard drives and digital audio tapes from the '90s and '00s are in greater danger than professional analog tapes from the '70s or '80s, regardless of tape formulation.
.

Kares said:

Militant said:

That actually makes a lot of sense when you consider that the first two albums were done in pro studios.

Prince blew his three album budget on mostly the first album and then the second. That's why Dirty Mind sonically sounds muddy - but given the style of the songs, it worked. It was recorded in a low budget leaky basement in his house.

So it's likely the stuff recorded at his house, between I guess 1980 and around 82 or so when he mostly started working at Sunset Sound (and before 1999 became a hit and he had more money to work with) where lesser quality recording material was used.

I suspect the most obvious example we have so far of this, is the version of "Wouldn't You Love To Love Me?" that ended up on Originals. There was only so much they could do with what they had, and thus it's sonically worse than the rest of the album.

.
WYLTLM sounds bad because it was likely sourced from compact cassette and it's drenched in reverb (it almost sounds like it was put through a spring reverb, lol), not because it was tracked on a lesser quality 2" tape. Howe & Co. probably either couldn't find the multitrack (maybe it's not marked on its box) or it was badly damaged and they couldn't save it.
.
Generally speaking though: hard drives and digital audio tapes from the '90s and '00s are in greater danger than professional analog tapes from the '70s or '80s, regardless of tape formulation.
.

Is there a way to artificially un-reverb this shit? It's such a good song.

LoveGalore said:

Question, though - When Prince would submit those progress tapes to WB, was he submitting actual 1/2 or 1/4 studio reels or was he submitting cassettes or DATs to them? The stuff off PR Deluxe was sourced from those progress submissions, I thought (with exception of Electric Intercourse?), and I can't imagine they just preserved cassettes or even DATs so well for 30 years.

.
I don't think he was regularly submitting professional reels, only cassettes. DAT machines didn't actually come out before the very end of the '80s ("thanks" to the idiots at the top of the major labels). And yes, analog cassettes preserve playability far better than DATs, even though they sound worse to begin with. But there's nothing worse than a DAT refusing to track properly and exhibiting serious drop-outs. Analog tapes are far less needy.

.

LoveGalore said:

Kares said:

.
WYLTLM sounds bad because it was likely sourced from compact cassette and it's drenched in reverb (it almost sounds like it was put through a spring reverb, lol), not because it was tracked on a lesser quality 2" tape. Howe & Co. probably either couldn't find the multitrack (maybe it's not marked on its box) or it was badly damaged and they couldn't save it.
.
Generally speaking though: hard drives and digital audio tapes from the '90s and '00s are in greater danger than professional analog tapes from the '70s or '80s, regardless of tape formulation.
.

Is there a way to artificially un-reverb this shit? It's such a good song.

.
I wish! But there isn't, unfortunately.

Kares said:

Wow, then that is amazing to me. In particular, "Katrina's Paper Dolls" has so much range and is so crisp and full bodied. Shocks me that it really came from a cassette tape, even with Grundman's mastering work done on it.

Also, since you're well versed in this stuff: When Howe says he is replicating a mix... I don't get that. So they have multi track reels and then a reference cassette as well? So for Don't Let Him Fool Ya, in that tape box pictured is just a studio reel with all the various tracks but they aren't actually collected into a song or what? What are they using to replicate a given mix from?

And when this stuff is digitized - would they digitize everything, so a folder on some hard drive would have a file for each stem and then a file for the song mixed?

LoveGalore said:

Kares said:

.
I don't think he was regularly submitting professional reels, only cassettes. DAT machines didn't actually come out before the very end of the '80s ("thanks" to the idiots at the top of the major labels). And yes, analog cassettes preserve playability far better than DATs, even though they sound worse to begin with. But there's nothing worse than a DAT refusing to track properly and exhibiting serious drop-outs. Analog tapes are far less needy.

.

Wow, then that is amazing to me. In particular, "Katrina's Paper Dolls" has so much range and is so crisp and full bodied. Shocks me that it really came from a cassette tape, even with Grundman's mastering work done on it.

Also, since you're well versed in this stuff: When Howe says he is replicating a mix... I don't get that. So they have multi track reels and then a reference cassette as well? So for Don't Let Him Fool Ya, in that tape box pictured is just a studio reel with all the various tracks but they aren't actually collected into a song or what? What are they using to replicate a given mix from?

And when this stuff is digitized - would they digitize everything, so a folder on some hard drive would have a file for each stem and then a file for the song mixed?

.

Noise reduction and audio restoration software are pretty sophisticated these days and can do a lot more magic without severe side effects than 10 or more years ago, so yes, an old cassette can be brought back to life – but of course it still won't sound as good as professional tape can.
.
On the 2" multitrack tape (that's the box pictured that just came out) you have 16 or 24 different recordings running parallel and separated by gaps to avoid cross-track leakage. (DLHFY is probably a 16-track recording.) So these separate tracks can be played back together so you can hear the entire song, but they are still physically separated on the tape and you have to adjust their levels to create a mix.
If they have a rough mix on a cassette that P or his engineer made back in the day, they can mix the multitracks to match the balance of instruments and parts on the cassette, and they can apply effects such as compressors, de-essers, EQs, reverbs etc to make it sound like the mix on the cassette (minus the tape hiss, distortion, wow&flutter and other issues, of course).
.
Transferring a tape to archival formats does mean digitising all the contents of the tape (all the different tracks - or stems - if it's a multitrack, including the calibration tones and unused elements) as well as extensive technical notes on the physical condition of the tape (the required bias settings, any speed-calibration info if needed, any damages, splices - if it was physically edited - etc etc...), detailed info on the equipment and settings used for the transfer, plus photos of all sides of the original boxes and the reel etc – and yes, everything goes into a folder.
.

* I've uploaded a copy of an average archive log for you onto the third sheet of my Paisley Park Vault spreadsheet (see link in my signature) so you can see what sort of info is usually saved during a tape transfer.*
.

[Edited 11/11/19 9:01am]

Kares said:

Whoa - never seen one of these before. Thank you! Really fascinating info.

Do we know how recent Prince was using analog reels? I seem to recall that he never actually stopped using them altogether despite also working with primarily digital mediums as well.

Like I don't expect there to be reels of Art Official Age but surely he was still working on analog tape in the early to mid 90s and not yet fully on hard drives?

LoveGalore said:

Kares said:

.

Noise reduction and audio restoration software are pretty sophisticated these days and can do a lot more magic without severe side effects than 10 or more years ago, so yes, an old cassette can be brought back to life – but of course it still won't sound as good as professional tape can.
.
On the 2" multitrack tape (that's the box pictured that just came out) you have 16 or 24 different recordings running parallel and separated by gaps to avoid cross-track leakage. (DLHFY is probably a 16-track recording.) So these separate tracks can be played back together so you can hear the entire song, but they are still physically separated on the tape and you have to adjust their levels to create a mix.
If they have a rough mix on a cassette that P or his engineer made back in the day, they can mix the multitracks to match the balance of instruments and parts on the cassette, and they can apply effects such as compressors, de-essers, EQs, reverbs etc to make it sound like the mix on the cassette (minus the tape hiss, distortion, wow&flutter and other issues, of course).
.
Transferring a tape to archival formats does mean digitising all the contents of the tape (all the different tracks - or stems - if it's a multitrack, including the calibration tones and unused elements) as well as extensive technical notes on the physical condition of the tape (the required bias settings, any speed-calibration info if needed, any damages, splices - if it was physically edited - etc etc...), detailed info on the equipment and settings used for the transfer, plus photos of all sides of the original boxes and the reel etc – and yes, everything goes into a folder.
.

* I've uploaded a copy of an average archive log for you onto the third sheet of my Paisley Park Vault spreadsheet (see link in my signature) so you can see what sort of info is usually saved during a tape transfer.*
.

[Edited 11/11/19 9:01am]

Whoa - never seen one of these before. Thank you! Really fascinating info. Do we know how recent Prince was using analog reels? I seem to recall that he never actually stopped using them altogether despite also working with primarily digital mediums as well. Like I don't expect there to be reels of Art Official Age but surely he was still working on analog tape in the early to mid 90s and not yet fully on hard drives?

.
You're welcome.
.
P was recording primarily on analog tape right until the very end, with phases of recording on digital tape (DASH) in the second half of the '90s and using a DAW (ProTools) later.
.
I guess most of AOA was done in ProTools as I remember seeing an Apogee Quartet interface and a Lacie hard drive lying around on the SSL board in Studio A during an interview with Josh, but a lot of HnR songs were tracked on 2" analog tape.
.

LoveGalore said:

- would they digitize everything, so a folder on some hard drive would have a file for each stem and then a file for the song mixed?

.
forgot to add earlier:
many professional archive facilities actually use tape storage formats (LTO, for example, which is probably the best imo), not hard drives. Tape storage has many advantages over hard drives for archiving purposes and LTO continues to evolve.

.

Speaking to what Kares posted above, I came across an interesting interview podcast with Bernie Grundman

http://www.thevinylguide....-mastering who, aside from doing a lot of Prince`s WB era mastering work was also charged with preserving Blue note`s archive, and his answer to the best preservation format surprised me. He was asked about various formats , and his preference was for analogue tape!

He argued (if I understood correctly) that a lot of digital formats were susceptible to hardware failures, and digital glitches in a way that high quality analogue masters were not. A very interesting interview if you can spare the time and forgive the intrusive commercials

andrewm7 said:

Speaking to what Kares posted above, I came across an interesting interview podcast with Bernie Grundman

http://www.thevinylguide....-mastering who, aside from doing a lot of Prince`s WB era mastering work was also charged with preserving Blue note`s archive, and his answer to the best preservation format surprised me. He was asked about various formats , and his preference was for analogue tape!

He argued (if I understood correctly) that a lot of digital formats were susceptible to hardware failures, and digital glitches in a way that high quality analogue masters were not. A very interesting interview if you can spare the time and forgive the intrusive commercials

.
I understand where he's coming from but IMHO as every format has its problems and none of them is absolutely future-proof (yes, analog tape itself has a fairly long shelf life, but properly maintained machines to play those tapes might be hard to find in 20 years), it is best to keep digital safety copies on at least two different archival formats (LTO and disk array) and keep one or two analog tape copies as well, preferably at different geographic locations.

.

What's also important is that no matter which archival technology you use, it'll have to be maintained and checked regularly. I know of several studios who thought it is safe to just back up projects on hard drives and lock them away – only to find out 10 years later, when they pulled out something for remastering, that the HDDs won't even spin up anymore.

.

Kares said:

Couldn't the digital formats be uploaded to a cloud and thus made into redundant copies forever?

LoveGalore said:

Kares said:

.
I understand where he's coming from but IMHO as every format has its problems and none of them is absolutely future-proof (yes, analog tape itself has a fairly long shelf life, but properly maintained machines to play those tapes might be hard to find in 20 years), it is best to keep digital safety copies on at least two different archival formats (LTO and disk array) and keep one or two analog tape copies as well, preferably at different geographic locations.

.

What's also important is that no matter which archival technology you use, it'll have to be maintained and checked regularly. I know of several studios who thought it is safe to just back up projects on hard drives and lock them away – only to find out 10 years later, when they pulled out something for remastering, that the HDDs won't even spin up anymore.

.

Couldn't the digital formats be uploaded to a cloud and thus made into redundant copies forever?

.
"The Cloud" is just a way of transportation: instead of sending your hard drives by DHL or whatever, you're sending your data via the cloud to an archive facility that is (most probably) using tape based (or disk-based) storage.
The real question is how much you can trust that company who runs that facility. How you send your data to them is secondary.

.
But you really don't want to send extremely valuable, unpublished and sensitive data over the internet and you also don't want the data servers storing your precious archive to be connected to the internet, so using the cloud is a bad idea for several reasons.
.

[Edited 11/12/19 5:27am]

Kares said:

On the 2" multitrack tape (that's the box pictured that just came out) you have 16 or 24 different recordings running parallel and separated by gaps to avoid cross-track leakage. (DLHFY is probably a 16-track recording.) So these separate tracks can be played back together so you can hear the entire song, but they are still physically separated on the tape and you have to adjust their levels to create a mix. If they have a rough mix on a cassette that P or his engineer made back in the day, they can mix the multitracks to match the balance of instruments and parts on the cassette, and they can apply effects such as compressors, de-essers, EQs, reverbs etc to make it sound like the mix on the cassette (minus the tape hiss, distortion, wow&flutter and other issues, of course).

.

Just a quick (out of topic) question for you, Kares. Some people are now using machine learning / deep learning for source separation (i.e. separating a stereo recording into a vocal track, drums track, etc.). Deezer has just released an open source project (Spleeter) that does just that, with pre-trained models. It works relatively well for vocals. Not much for the rest (bass, drums, piano), according to my tests.

.

With thousands of multitracks / stereo mixdowns pairs, do you think it would be possible to train a model how to mix a multitrack automatically given a reference stereo mixdown? The model could be as simple (i.e. only a balance of the different tracks) or complex (e.g. compressors, reverbs, etc.) as needed. Do you know people working on that? Do you know if there are already software solutions to do that?

.

Reading/hearing Michael Howe and Niko Bolas talk about the mixing process, I've always thought that, in an ideal world, the mixing process should be as automated as possible. We just know what the result is when you have a human being in the equation (e.g. way too much reverb, etc.).

[Edited 11/12/19 5:30am]

olb99 said:

Kares said:

On the 2" multitrack tape (that's the box pictured that just came out) you have 16 or 24 different recordings running parallel and separated by gaps to avoid cross-track leakage. (DLHFY is probably a 16-track recording.) So these separate tracks can be played back together so you can hear the entire song, but they are still physically separated on the tape and you have to adjust their levels to create a mix. If they have a rough mix on a cassette that P or his engineer made back in the day, they can mix the multitracks to match the balance of instruments and parts on the cassette, and they can apply effects such as compressors, de-essers, EQs, reverbs etc to make it sound like the mix on the cassette (minus the tape hiss, distortion, wow&flutter and other issues, of course).

.

Just a quick (out of topic) question for you, Kares. Some people are now using machine learning / deep learning for source separation (i.e. separating a stereo recording into a vocal track, drums track, etc.). Deezer has just released an open source project (Spleeter) that does just that, with pre-trained models. It works relatively well for vocals. Not much for the rest (bass, drums, piano), according to my tests.

.

With thousands of multitracks / stereo mixdowns pairs, do you think it would be possible to train a model how to mix a multitrack automatically given a reference stereo mixdown? The model could be as simple (i.e. only a balance of the different tracks) or complex (e.g. compressors, reverbs, etc.) as needed. Do you know people working on that? Do you know if there are already software solutions to do that?

.

Reading/hearing Michael Howe and Niko Bolas talk about the mixing process, I've always thought that, in an ideal world, the mixing process should be as automated as possible. We just know what the result is when you have a human being in the equation (e.g. way too much reverb, etc.).

[Edited 11/12/19 5:30am]

.
Separation of instruments/parts of a recording is now a reality indeed, although they are still at the early stages of development and afaik currently it only works on recordings with only a few parts. (Ravi Shankar's Woodstock festival concert has just been restored and remixed this way for the complete Woodstock box set that came out recently.)
.
I'm sure there will be software for creating automated mixes and it will be able to learn the certain sound of another mix to create something similar. I don't know if it's available yet but I'm sure it will be in the near future.
HOWEVER: I admit I hate the thought of that. Mixing is largely a creative, artistic process. I can imagine that producers of K-pop or any other assembly-line manufactured pop music will use such quick solutions, but it would be blasphemy to try to apply that to a real artist's (such as Prince's) work. There's no way an AI-based mixing software will come up with ideas such as "hey, let's just mute that bass track, it'll freak out the record execs!". It would kill originality. It would kill Music.
.
Just try to imagine trusting AI to edit and mix something like 'Bitches Brew'. I can't. What Teo Macero has done with that stuff, all that editing in post-production, is equal to composing. He was given the go ahead by Miles to create something new from the session tapes and he did.
I'd never want to eliminate the human element in creating music and that goes for mixing too.
.

[Edited 11/12/19 6:01am]

IstenSzek said:

forgive me for asking as i'm sure this has been discussed multiple times but there is simply
so much information these last few years and i'm somehow still clueless as to who will get
to release post WB material and from what date in the future? 2020? 2021?

The press release only mentions that Sony got the licensing rights for 35 previously released albums.

So i wouldn't be surprised if Sony had to pay extra $$$ to get the Vault rights by 2021 or so.

It's all there : https://www.rollingstone....ms-666992/

BartVanHemelen said:

olb99 said:

.

Exactly. I can confirm what Niko Bolas said. I work on an audiovisual archive containing more than 50 years of recordings. I would be far more worried about files/songs stored on harddrives from the 90s/00s than on tapes from the 70s/80s...

[Edited 11/9/19 2:44am]

.

I wouldn't be surprised if they've done some form of triage, and determined that while perhaps the 90s/00s material is in similar danger as the older material, it is not a priority from an artistic and money-making point of view.

as you would often reply...stop making shit up!

Militant said:

I was under NDA to promote the recent reissues, so I might be able to shed some light based on discussions had. And I will be careful here just so I don't piss anyone off. Some of this is still my speculation based on fragments and my own experience being in the music industry for nearly 20 years. So take this with a grain of salt - I'm not presenting all of this as gospel facts.

I worked on a list of possible ideas around the Emancipation re-release specifically, and during that time, I did raise the subject of outtakes, and questioned whether it'd be possible to include Emancipation era outtakes.

My understanding is that Sony do have the rights, but not the green light from the Estate right now. Archival of what's in the vault is a long process, and as much as we'd all love to have his job, Michael Howe's task is actually quite arduous. For perhaps obvious reasons - including the age of the material and the business need - there is still archival work happening around things that are contained on physical reels. Preservation, etc.

By the time Prince left Warners, we're largely in the digital age. So you're talking files now - not reels. Although, there have stilll been many instances where stuff has been mixed onto tape in analog. But those aren't the only versions of these songs in existence. So you're not talking extinction level event, where if an old reel is fucked, the song is wiped from existence in a practical sense.

There is an urgent need to fix and preserve the oldest material in the vault. There are old reels that are degrading, and the oldest material is at the most risk. I heard directly from Manuela that even during their marriage, there were instances where old tapes were brought up from the vault and were already damaged at that time - and that was 15 years ago. That's not to say that things can't be fixed - they often can. But action needs to be taken as soon as possible to do that, and that's what's happening at Iron Mountain.

So - the focus is on that. And on a business level, we must all accept that 70's and 80's Prince is of more worth to the general public than the post-WB stuff. For these two reasons, that's where the archival work is happening, and other things are on the backburner for now.

That's not to say that there aren't post-WB projects that need re-issuing. And so that's happening. But for the near future at least, these will be straight re-issues, like what we've seen with Emancipation, Chaos, etc.

really valuable info, thank you. Personally I hope his latter tracks / albums are released next such as Black is the new Black. P&M Paisley shows etc.

fredmagnus said:

IstenSzek said:

forgive me for asking as i'm sure this has been discussed multiple times but there is simply
so much information these last few years and i'm somehow still clueless as to who will get
to release post WB material and from what date in the future? 2020? 2021?

The press release only mentions that Sony got the licensing rights for 35 previously released albums.

So i wouldn't be surprised if Sony had to pay extra $$$ to get the Vault rights by 2021 or so.

It's all there : https://www.rollingstone....ms-666992/

thanks fredmagnus

Kares said:

olb99 said:

.

Just a quick (out of topic) question for you, Kares. Some people are now using machine learning / deep learning for source separation (i.e. separating a stereo recording into a vocal track, drums track, etc.). Deezer has just released an open source project (Spleeter) that does just that, with pre-trained models. It works relatively well for vocals. Not much for the rest (bass, drums, piano), according to my tests.

.

With thousands of multitracks / stereo mixdowns pairs, do you think it would be possible to train a model how to mix a multitrack automatically given a reference stereo mixdown? The model could be as simple (i.e. only a balance of the different tracks) or complex (e.g. compressors, reverbs, etc.) as needed. Do you know people working on that? Do you know if there are already software solutions to do that?

.

Reading/hearing Michael Howe and Niko Bolas talk about the mixing process, I've always thought that, in an ideal world, the mixing process should be as automated as possible. We just know what the result is when you have a human being in the equation (e.g. way too much reverb, etc.).

[Edited 11/12/19 5:30am]

.
Separation of instruments/parts of a recording is now a reality indeed, although they are still at the early stages of development and afaik currently it only works on recordings with only a few parts. (Ravi Shankar's Woodstock festival concert has just been restored and remixed this way for the complete Woodstock box set that came out recently.)
.
I'm sure there will be software for creating automated mixes and it will be able to learn the certain sound of another mix to create something similar. I don't know if it's available yet but I'm sure it will be in the near future.
HOWEVER: I admit I hate the thought of that. Mixing is largely a creative, artistic process. I can imagine that producers of K-pop or any other assembly-line manufactured pop music will use such quick solutions, but it would be blasphemy to try to apply that to a real artist's (such as Prince's) work. There's no way an AI-based mixing software will come up with ideas such as "hey, let's just mute that bass track, it'll freak out the record execs!". It would kill originality. It would kill Music.
.
Just try to imagine trusting AI to edit and mix something like 'Bitches Brew'. I can't. What Teo Macero has done with that stuff, all that editing in post-production, is equal to composing. He was given the go ahead by Miles to create something new from the session tapes and he did.
I'd never want to eliminate the human element in creating music and that goes for mixing too.
.

[Edited 11/12/19 6:01am]

.

Thanks for your answer. I agree that the mixing process is a creative process. What I had in mind was a very specific/narrow use of that technology to produce a basic mix from a raw multitrack recording and a reference mixdown. That basic mix could later be reworked by a mixing engineer, of course. The use cases I have in mind is the the Prince archive and the work that Michael Howe and Niko Bolas had to do to create "releasable" material, but also our own archive where I work, where we basically have thousands of multitrack recordings and stereo mixes (from live concerts), but not way to mix/remix the multitracks to get the stereo mixes. The mixing process was done live, during the concerts, and we have no way to recreate exactly what was done at that time... (I'm a software engineer, so sorry if I sound a bit out of my element. )

olb99 said:

Kares said:

.
Separation of instruments/parts of a recording is now a reality indeed, although they are still at the early stages of development and afaik currently it only works on recordings with only a few parts. (Ravi Shankar's Woodstock festival concert has just been restored and remixed this way for the complete Woodstock box set that came out recently.)
.
I'm sure there will be software for creating automated mixes and it will be able to learn the certain sound of another mix to create something similar. I don't know if it's available yet but I'm sure it will be in the near future.
HOWEVER: I admit I hate the thought of that. Mixing is largely a creative, artistic process. I can imagine that producers of K-pop or any other assembly-line manufactured pop music will use such quick solutions, but it would be blasphemy to try to apply that to a real artist's (such as Prince's) work. There's no way an AI-based mixing software will come up with ideas such as "hey, let's just mute that bass track, it'll freak out the record execs!". It would kill originality. It would kill Music.
.
Just try to imagine trusting AI to edit and mix something like 'Bitches Brew'. I can't. What Teo Macero has done with that stuff, all that editing in post-production, is equal to composing. He was given the go ahead by Miles to create something new from the session tapes and he did.
I'd never want to eliminate the human element in creating music and that goes for mixing too.
.

[Edited 11/12/19 6:01am]

.

Thanks for your answer. I agree that the mixing process is a creative process. What I had in mind was a very specific/narrow use of that technology to produce a basic mix from a raw multitrack recording and a reference mixdown. That basic mix could later be reworked by a mixing engineer, of course. The use cases I have in mind is the the Prince archive and the work that Michael Howe and Niko Bolas had to do to create "releasable" material, but also our own archive where I work, where we basically have thousands of multitrack recordings and stereo mixes (from live concerts), but not way to mix/remix the multitracks to get the stereo mixes. The mixing process was done live, during the concerts, and we have no way to recreate exactly what was done at that time... (I'm a software engineer, so sorry if I sound a bit out of my element. )

.
I see – yes, with such an archive of live recordings I understand the need to create quick, draft mixes in bulk. Unfortunately though I don't know of any software solutions for that. And even if it exists I'm sure it'd only be good enough to create very-very basic draft mixes.
.
On the other hand, let me know if you need a hand in restoring/mixing/mastering anything

.