Can My Damaged Recordings be Repaired?

Much of my source material (uncompressed CD track rips; no SACDs or vinyl), are 60s pop, r&b and soundtracks, which were often victims of excessively applied compression. And some tracks during multi-vocal passages and when orchestrations get busy sound “congested” (though not necessarily from waveform clipping due to record gain overload). I presume it was likely due to poor miking/baffling (??), other acoustical causes; others say it’s the result of what must have been accidental tape saturation.

Of course, today the only time you hear the term used it’s always about how great tape saturation sounds when its deliberately added to a recording! YECCH, what perversion!

So, while not outright crappy (??), much of my music was certainly less than pristinely recorded and/or mastered, even though almost all were issued by major labels. And they may sound especially poor over the new custom built high res horn type speakers I will soon be installing.

When will there be algorithms or AI solutions which can fix treasured vintage recordings with this kind of damage?

Hello, Welcome…

A digitized analog recording generally will have ‘limiting’ in the ADC (Analog to Digital Conversion) to prevent digital clipping… In a multi-track recording, compression may or may not have been applied (however we can be safe in presuming creative compression has been used in tracking for the genre you are describing for the most part)

Creative compression cannot be, and should not be, monkey’d with as this has been done in the creative artistic decisions during production… However, it is a far stretch to presume the digital-audio products you are playing in Audirvana, have been compressed in ADC mastering, unless this compression is inherent in the analog master ‘source’… If this compression exists in the source master recording used to produce the digital-audio master recording, the digital-audio iteration is a transparent view into the recording, and it-is-what-it-is… Trying to re-think the creative process of a given analog or digital production is futile… What is happening in the high-resolution space, is record companies are re-mastering popular and historically significant productions… a good example is Giles Martin re-mixing and mastering Beatles recordings that have been digitally archived or from original master tapes. This is what you want to see… Any post-production attempt (by you) to change the creative decisions made in the production and mastering digitization processes, will go unrequited.

:notes: :eye: :headphones: :eye: :notes:

Thanks for your reply. I’m no expert with any of this, and without software and/or hardware tools that strongly lend themselves to being utilized for making audible improvements along the lines discussed, there’s nothing I could or would likely attempt. Otherwise, any desired wish that these analog recordings had been hit with much less analog compression than they were-however “creatively” (judiciously) applied and/or or as of necessity during their digitization, stems from much from what they will likely sound like on my new speakers, driver selection for which are still ongoing.

Meanwhile, my friend Pierre generously spent three hours evaluating 50 + WAV files I had sent him in a zip file. Pierre’s test bed: mac mini; Audirvana upsampling to 176.4kHz; Holo May DAC: Aikido line stage; VFET power amp feeding his speakers; posts 15,266, 15,276.

As these are two-way > 94db/w/m speakers with precisely EQed beryllium horn drivers, it’s understandable why Pierre readily found many of those recordings to be lacking in dynamic range, and thus-as would most listeners?-found them undesirably compressed. Thus, it would seem that when at least some of those multitrack recordings were originally mastered during the final mix down to stereo (or mono)-long before being digitized-that compression was applied not only to low frequency content but to all tracks to ensure prevention of exceeding the dynamic range of vinyl. Of course, as today’s digital formats can capture true orchestral hall levels was why I sought a way to undo some of this globally applied compression originally applied to those recordings.

But as mentioned earlier, with most of these recordings, Pierre also found them to sound “congested”. How much do you suppose that’s due to applied analog mastering compression and/or unintended tape saturation? Certainly, the presence of that was hardly and “artistic” choice. These files are available via my Google drive for anyone to observe how this artifact detracts from the enjoyment of these otherwise reasonably good quality vintage recordings.

If compression is imbued in an original ‘mother’ master recording… There is no way, to restore the dynamic-range that has been reduced in the original master source… You would need to get your hands on the original 2-track or 4-track master tape, or vinyl/acetate/wax ‘mother’ recording and re-master the production… There are good recordings and bad recording and very good recordings derived from analog masters… 2xHD is a label that exemplifies how well the re-mastering process can be done… They produce high-resolution products from analog masters from the 50’s and 60’s and digital productions that are state-of-the-art and reveal every nuance of a performance.
https://2xhd.com/high-res-home.html

It is a mistake to assume that pre-digital recordings can reproduce the dynamic-range and frequency response of modern digital recordings done at high sample-rate PCM 24 bit or DSD…

To question the production techniques of any recording is just a subjective interpretation, beholding to cognitive biases that have one wishing it had been done differently, for some cognitive rationale. It’s like interpreting Rodin’s ‘The Thinker’ and wishing he had set the subject in a different pose…

If you are describing MP3 and other lossy audio products… well good luck trying to bring these to full resolution… up-sampling and some psychoacoustic processing can put some life back into these products, but they will never match-up to a high-resolution source master recording of the same production.

:notes: :eye: :headphones: :eye: :notes:

FWIW, all of my recordings are commercially pressed CDs from original labels, not mp3s.

Not surprisingly, the catalog at that website contains none of the titles of particular interest. As for the original analog masters, while some may still exist in the (Capitol, Warners, Sony, et al) vaults, I would hardly have the facilities to remaster them, much less the permission to ever borrow them. And the rest were most were likely doomed, however accidentally.

So, while thankful for what I have it’s sad that there’s likely no fix for their condition. At least Pierre, my builder and I believe that they will still sound their best on my new system.

Maybe something here:

Audio restoration - There’s An AI For That (theresanaiforthat.com)

Yes, I bumped into that hit this week while doing a related search.
Forgot to check it out. Thanks.

The fix is generally found in the perspective one takes on recordings of any genre and time-period… Nobody, can go back and re-produce a collection of tracks done in a particular contextual time and space on the historical time-line of commercially based music productions… Nothing is in stasis… This is why they are called ‘recordings’, as these represent a particular snap-shot of artistic and technological decisions, relative to that time and space… It just so happens that with a well archived digitized iteration of these pieces of art, those recordings can be auditioned through very transparent playback systems today… Music appreciation and the appreciation of artistic decisions made in the production processes is beholding to subjective interpretations, intrinsically tied to cognitive biases, as you clearly point out, regarding your appreciation of the 2xHD catalog of digitized analog productions… So, what does the term “condition” really mean in the context of all the potentials for appreciation of a given recording, or any given music product on any given playback system?

Your system will produce a quality of sound that is intrinsic to a myriad of variables and potential influences, that shape the final audition of any music playback on your amalgamation of components and tweaks and acoustic environment in which your speaker system is installed.

Your best approach is to make your playback system as transparent to the source signal(s) as possible within your means… Only you will be the arbiter of this experience. When you have accomplished this, you will then, need to come to grip the reality presented in the music productions that you listen to… nothing more, nothing less.… Some recordings are good, some bad and some are great… this is the reality. :sunglasses:

:notes: :eye: :headphones: :eye: :notes:

You might check out these… https://gearspace.com/board/best-studio-gear/1232179-ten-best-audio-repair-software-amp-plug-ins-2023-a.html

iZotope has a good reputation… You might consider contacting them to see if they might help you try a few samples of your problematic tracks with their software… I think you could ask the same of all the vendors on this list…

You might look at hardware or software compressors/expanders which often have expanders which might do the trick… Not 100% effective at reviving dynamics… Some may be better than none…

No reason not to ask around a few pro studios too…

Actually, while I have an old version of Izoptope Rx Advanced I haven’t had time to use it much. Happily, I’ve only had rare occasion to use its De-Clip utility. https://www.izotope.com/en/products/rx/features/de-clip.html

Again, as mentioned, clipping (record gain overload) is not primarily what afflicts most of my troubled recordings, at least as far as I can audibly detect. The main problems are lost dynamic range due to the amount of compression applied to individual tracks and/or the final mono or stereo mixdown of these mostly vintage 60s recordings. The other very noticeable problem is congested sound during multi-vocal and orchestral passages. I wouldn’t know whether this latter damage is due to the amount of compression, from poor miking/baffling techniques and/or other acoustical or electronics mechanisms.

But I found nothing in my Rx5 Advanced tool set that could be used to undo any of this damage. And when now searched [ izotope rx “congested” ] there’s no hint that the current version of Rx Advanced can do so either.

As for Izotope tech support I can tell you that it’s not what it used to be, at least since the company merged with Native Instruments. Izotope developers no longer give users opportunities to submit feedback, such as requesting new features. And there never was an official community forum, as there is with most DAW platforms and other audio software devices. Last week I did send a support request via https://support.izotope.com/hc/en-us/sections/6658006518033 describing what I was hoping to be able to do in Rx; no reply thus far.

Sadly, given the fact that atrocities like the Loudness Wars still rage on, how serious would developers of these tools even consider these problems to be damages worth addressing.

But to quote a fellow member at diyaudio.com:

AI bends the rules that we have accepted for so long. I do anticipate a time when someone can simply install some software and run their old, DR-compressed (or badly mastered) material through it and end up with something much, much better.

I’ve recently installed and tested an application named Upscayl, which is a free and open source software that uses AI to upscale photo/picture resolution. The results are nothing short of incredible. It can take a low resolution JPEG that is completely inadequate for printing and upscale it to a gorgeous, printable photo. I do believe it is only a matter of time before we have software for music that is this impressive.

And since there’s apparently no other way to send user feedback to Audirvana, my reason for joining this forum was to express my feelings about what features a subscriber accessible audiophile player should have by now, or that its developers are now hard at work to deliver: AI-based restoration engines that can make my troubled recordings a delight to behold over the high res system I’m building.

With all the promise and trepidation that AI presents to humanity, this would certainly be one application where everyone wins.

1 Like

The closest possible tool will be the ‘Rebalance’ module in RX10…

It sounds to me that you have not come to grip the reality and have not learned to appreciate the warts and wrinkles of the artistic endeavors that you hubristically assert as being ‘damaged’ recordings… Such a myopic perspective… :smirk: I have several CD rips from that era and genre that when modulated to DSD128 reveal the true essence of the artistic intentions, with extremely focused clarity of contextual harmonics, contextual dynamics and contextual spatial relationships of the performance(s)… It sounds like you have been reading too many MIX articles and other urban myths surrounding recording-arts in the 50’s and 60’s… Are you going to tell me that the intentional distortion that Jimmy Page got from inputting guitar direct into the Helios desk in Olympic Studios on some tracks, is a problem? … or … Steve Miller never produced a great sounding record? Or Eddie Kramer, sucked as an engineer? and Bob Ludwig was a hack? Barry Gordy, Ray Charles and Amet Ertegun, George Martin with Geoff Emory produced lousy sounding records? … the list rolls on… :roll_eyes: Your perspective is myopic at best.

Let AI produce new music… Leave the past alone and learn to appreciate the warts and wrinkles of those recordings you think are damaged… Soon enough, music will be rationalized to the lowest common-denominator that the algorithm finds most profitable and ubiquitously appealing.

:notes: :eye: :headphones: :eye: :notes:

1 Like

True words of wisdom…