Hi, I’m just revisiting a live studio recording session from 25 years ago that was instantateously mixed down to a stereo pair and published to a TDK SA!!
This “master” has only been played to duplicate a couple of masters then it was stored and now I have imported it into Cubase. There was some ghosting of magnetic imprint from adjacent spool wrap and a few minor drop outs but that’s only obvious if you know it’s there.
I have tweaked it to get a more interesting stereo spatial in important frequency bands and am sort of pleased with the digitally rehacked result. But… I can’t but help wonder if there’s any other tech out there that can “reverse engineer” a mix back to constituent tracks. Oui les cochons comme oiseax. I know that’s a stab in the dark but if one can pick out salient frequencies then intelligently analyse which could or could not logically come from a defined instrument then that’s sorted. Does this exist?
XX
Cant say you can reverse engineer, but you can bake the tape, does wonders for restoring some quality, removes the stickiness from being wound tight during years of decay.
My former studio partner (Don Sklepowich) was also a restoration specialist and archivist with the University of Toronto Robarts Media Commons, he could have advised nicely on this, sadly he passed away two years ago, we collaborated almost 30 years ago on various recordings to half and one inch tape, he rejuvenated some of those old items nicely. He died before we could finish. 
Have a look at
Might help you get a better dump to digital, there is some risk of course.
Here’s Don posing during a class he taught on the subject. He was well liked at the university, and to many an inspiration.
No. XX
@Northwood
Thank you for sharing your fond memories - love that hat - bless him.
I’ll give those links a visit. I had wondered about importing from the master again. It’s only on cassette and not wide tape. I did use a quality tape deck but maybe there are better machines to do the job.
XX
this might help now that you’ve been more specific
or have someone else do it…
I still have some old cassettes that I have not transferred… hmmmm, better get on with it. 
No, it does not. It never will exist.
You can’t “unbake” a cake; you can’t get the constituent ingredients back to the state they were in inside the mixing bowl.
Hi thanks for the answers and for the links.
I have contacted Mark re his audio restorations and transfers. He has a good looking setup there. He is in the USA and I’m in UK. Just a bit worried about my tape flying such a long way. If he were just down the road I’d probably give it a go. It would be more handy to find someone in UK who comes as recommended as Mark. Then if it falls out of the postman’s sack I have a better chance of finding it.
I imported to Cubase with a clean NAD cassette deck and Lexicon Omega. Does anyone have a handle on the likely improvements in transfer quality using more professional kit or am I likely to have got it as good as I can get it by my DIY method?
Thank you all.
X
You seem to have good playback gear, so I would think its really down to the quality of the tape… so you probably would not see much improvement changing the hardware. Demagging the tape playback head and cleaning it is also a must. I used some pretty good equipment on my last transfers and still found there was some wow inherent… that’s down to the tape mechanism and the aging process. Tape stretches with use also. Its funny to think that people spend hundreds on plugins trying to emulate that now! 
Yeah maybe I’ve got the perfect trendy mix for the present day by mistake how cool’s that? 
.
On the subject of reverse engineering I don’t think we can say it will never be done. I am fully aware one cannot unbake a cake but we have not necessarily chemically altered the ingredients in this instance. They are still raw and in inseparably mixed up in the bowl prior to being cooked. But if I can manage to manually analyse and break down the mix into a fine resolution stack of frequency bands then go in there and chop out a few picoseconds of pop and boost those frequencies that more uniquely define the instrument requiring a boost then that’s part way there. I would not say that it will never be done. It just hasn’t happened yet. If an AI algorithm can learn and analyse then it will come. Maybe we wait for defence research into submarine signature analysis to filter down to the world of music. Or wait 'til we understand how bats communicate at such high frequencies of aural-brain perception then we’ll get it.
X
Submarine signature analysis could achieve the required end result by “filter(ing) down” for purposes of reverse engineering a song into it’s component instruments/parts. It would have to be made greatly more powerful. Understanding a four-piece ensemble of bat music would be similarly complex, requiring not only vast amounts of computing power, but also someone to translate bat-speak.
Like I said, it will never happen.
I would say your ears can easily hear the parts, so certainly your brain can reverse engineer the mixdown, and so it must be possible to program a computer to do it. But until that is successfully done, the easy route is to re-record the parts as you hear them (maybe even play them better), and mix it again. That’s what all the garage bands have been doing forever! You don’t even have to create a score, although that is not so hard. You just need to let your new musicians listen and learn. If what you have on tape is good, it should not be hard to find musicians willing to reverse engineer it and play it.
That’s not “reverse engineering”. It’s not engineering at all. It’s learning the song, playing it, and recording it.
Sometimes the moment that you capture something is only to be that moment captured. Let it live in the glory for all that was.
Precisely. In fact, the outcome is usually better than the original, provided all the participants are truly invested.
And to TEEF’s point, it is what it is, so really its about your choices here.
I’ve had some success recreating things I did 40 years ago and lost the tracks … its a great challenge. Its never going to be exactly the same, but seeing as technology has improved slightly since then, your results are generally going to be better, unless of course you’ve also lost 40 years of brain cells…
I think the closest you can get, and this depends a lot on how the original was mixed, is working with M/S separation tools such as what Waves offers.
It’s not what you are asking for, but I will assure you that sometimes, or perhaps the majority of times depending on the music, you can hear instruments and even vocals that you can’t hear in the master mix. I have done this with some hits I’m very familiar with and after M/S you get a different perspective…with certain hit songs.
A good example of breaking down a track and listening MS on separate channels Abbas Dancing Queen. You can hear a synth arrpegiator that is mostly or perhaps entirely covered up in the master. Of course if Michael Tretow had muted it, it would obviously be missing for him, but for the listener, it gets covered up.
Dancing Queen vocals. I’m not sure if it’s Frida or Agnetha, but immediately following the beginning piano roll you have their vocals doing ohh’s and ahh’s for 8 measures. Using M/S you will pull out one of the girls individual alto voices. In the master it’s impossible to hear but M/S it is distinctively clear.