Greetings all!

This question relates to audio so I'm starting out here. It doesn't relate to video or film production so if it should be moved to the "everything else" forum please do so.

Recently an old friend presented me with a box of family records that are 50 to 60 years old, apparently having some sentimental value. I've been asked to convert these to digital form (sounds easy!) but of course I want to do the best reasonable restoration job. So far the laminated plastic w/ cardboard center (packed as premiums in boxes of breakfast cereal) are the worst of the lot.

I've read a lot of discussions recommending and comparing e.g. Audition, iZotope, and a few other combinations. One product family that sounds interesting (especially as discussed by 78-RPM fans) is made by Algorithmix, which I believe is a German company. Specifically I'd like to try their product Sound Laundry. Latest version supposedly is 2.5. I found a special "Demo" version on their website; it installs on my machine, but hangs and crashes. There is also supposedly a "Full" version, which is slightly different; the manual says the Full will work in a similar demo mode if user name and security key (which I don't have) are not entered upon installation. I've emailed Algorithmix but have gotten no response.

Do anyone know whether Algorithmix is still a viable going concern? And can anyone suggest where I might get the "Full" version of Sound Laundry (obviously without the registration info) so that I can at least get it to work in demo mode and evaluate whether it will be useful with my present project?

Thanks to anybody who can help out with this mystery! Always a treat to login and catch up w/ you folks.

Regards,
Greg Miller

Hi Greg, iZotope RX Advanced (https://www.izotope.com/en/shop/rx-10-advanced.html) is favored by many pros, and can attenuate and/or remove many sonic annoyances, however it ain't cheap, btw, iZotope historically has very attractive sales during the holiday season.
Steinberg's SpectraLayers (https://www.steinberg.net/spectralayers/) cost less than iZ's RX, but does not have as many 'automatic' tools. Though SLP is amazing otherwise and can remove noises that were previously imposable to eliminate without serious artifacts. It can also separate instrument stems (or other content) from a mixed file (aka, unbake the cake)..
I do not have much experience with the free or low cost restoration tools.. Magix Sound Forge (https://www.magix.com/us/sem/sound-forge-pro/version-comparison/#c1555170) Audio Studio is a cost effective audio recording/editing DAW and includes some basic restoration tools. Sound Forge Pro has more/better restoration tools and the SF Pro 'Suite' version includes SpectaLayers, which IMHO is worth the upgrade cost alone. Sound Forge Cleaning Lab (https://www.magix.com/index.php?id=26126&L=52&AffiliateID=146&phash=e8Zp8lTUPcNlzrGv&pk_campaign=Brand_Sound_Forge_coreFX_US_%28G%29&pk_kwd=magix%20cleaning%20lab&pk_source=googleAds&pk_medium=cpc&pk_content=472338084674&gclid=Cj0KCQiA99ybBhD9ARIsALvZavW2nit74NxD1RInON6Ni9eJzUO8yRIbj6L4Kkw2HdywnLKV6yHA1AMaAmGuEALw_wcB) is also relatively low cost
From what I understand, the free Audacity (https://www.audacityteam.org/l) has been updated and has restoration tools as well.

Hi Rick, thank you for the detailed reply.

I've heard iZotope mentioned many times over the years. No doubt it's one of the best NR packages around. I'm not sure I could justify the price for what will probably be a one-time project, this one batch of family records for a good friend. I don't recall hearing anything about Spectral Layers.

Several years ago I tried a handful of low-cost NR programs, to process just a few of my own records. At the time I concluded that Cool Edit Pro (well, that sets the time frame, doesn't it?) was better than any of the cheapies. Of course that may have changed. I did, back then, find that Diamond Cut could do a few things a bit better than Cool Edit. I was on the DC beta list for a few generations, so that helped with the budget.

I asked particularly about the Algorithmix products for a reason. I learned about them on some other forums that talk a lot about vinyl, where they get a lot of positive comments. The Algorithmix line supposedly were developed with the goal of cleaning up LPs and 78s. In fact they apparently were bundled with a German brand of phono preamps, marketed as TerraTec. So I thought that, rather than try some generic NR packages, perhaps it would be worthwhile to start with something intended specifically for phono records. And I figured since we have a fair number of Brits, Aussies, and some Europeans on these forums, maybe someone would have some up-to-date info about the Algorithmix or TerraTec.

Anyway thanks for all your info. I may watch for a holiday sale at iZotope, and will try to learn a bit more about Spectral Layers, at the very least.

Thanks again,
Greg

Pro tip: check if someone has done it already for each particular record (and uploaded to YouTube etc). Saves a ton of time.

Andrew

Thanks, Andrew. Very practical tip, and it may save me some time in the long run. In this case, however, many are one-off acetates of family members recorded prior to the '50s. If only they had stored them without ever playing them, the job would be a lot easier!

A wash in soapy water (a drop of dishwashing detergent does wonders), plus a rinse and dry can be the best dust crackle-pop noise reduction possible. Even with regular LPs people are afraid to give the things a wash ... and they're made out of vinyl!

Andrew

Any time I'm transferring vinyl to digital, I wash it first, with a dilute solution of isopropyl in distilled water, and just a few drops of wetting agent.

Alcohol will damage shellac and acetate, though, so then it's just water & wetting agent.

There are two big problems with acetate. First, the acetate itself exudes a white oily substance which ends up coating the surface, and which will gum up the stylus, so it *must* be removed. But water and oil don't mix, so you're counting on a tiny bit of detergent for the chemical cleaning. Some folks recommend using ammonia to remove the oil, but cautiously. Worst of all, I've read horror stories of the vinyl layer actually becoming detached from the aluminum substrate, so cleaning has to be done very gently (and with much prayer).

Having said all that (in response to cleaning), the ticks and clicks from dirt are not the worst noise. Of course there are clicks and pops from surface damage (dragged needles, etc.) which can be removed. But the worst problem is the groove wear from repeated playings of the soft acetate, most likely with a steel stylus and tracking force approaching ounces. That creates an unending but not-constant "whoosh whoosh" with a level that approaches program level. Noise-sample-threshold NR doesn't fix this as well as I'd like. From what I've read, the Algorithmix suite is especially good with this kind of noise ... hence my specific quest.

Greg

When the usual restoration tools do not work satisfactorily, I turn to precise spectralgraphic editing with SpectraLayers Pro. The initial leaning curve is a bit steep, however if you're familiar with Photo Shop or other photo editing software, it is a little easier. I have SpectraLayers Pro, but there are lower cost. SpectraLayers versions (https://www.steinberg.net/spectralayers/compare-editions/). SL Pro has the full 'Unmix' option, the others can separate voice only from the other elements.. SLP can be used as a stand-alone application or an ARA plug-in..
iZotope's low cost RX 'Elements' version has the usual restoration tools which work well. but does not offer spectral editing. I am not sure about RX [i]Standard'/i] and RX Advanced. i$ too pricey for me. I am not sure if the RX Adv. trial version can be used w/o 'save' or trademark limitations.

Greg, you mentioned Cool Edit Pro, which became Adobe Audition. I have used Audition for some clean up of vinyl records transfers. It works well on low end rumble, hiss and small clicks. You have to go through each file and manually isolate some of the other clicks if you really want to clean it up, though. Since I already subscribe to the Adobe Creative Cloud and the transfers I've done have been for either myself or family, it did the job well enough for the situation.

Have fun!

I think it's about time Greg posted some .wav sample files for us to evaluate and try our audio repair skills on. :-)

Andrew

Recently an old friend presented me with a box of family records that are 50 to 60 years old, apparently having some sentimental value. I've been asked to convert these to digital form (sounds easy!) but of course I want to do the best reasonable restoration job...

Greg. If you don’t have a lot of experience in audio restoration and yet want an audio suite that has some pretty powerful tools, but does not require a steep learning curve, you may want to have a look at SOUND FORGE Audio Cleaning Lab 4. It’s a wizard and preset-based toolset, but within those wizards and presets, there is a lot of fine-tuning capability to chase down the results you are after. It also incorporates a fair number of iZotope's useful sound restoration plugins. Great for old records, well any old audio really

I’ll put this recommendation into some context. For the last few years, I’ve been doing a lot of audio restoration on a history documentary TV series for the government. A lot of the source material comes from our National Film and Sound Archive. This material encompasses old scratchy radio recordings, records, audiotapes, film newsreels and all manner of archive audio and video files. To clean all this up I’ve been using a combination of tools that include iZotope’s Ozone and Magix’s Sound Forge with a massive collection of VST Waves plugins, Antares’ Sound Soap and Celemony’s Capstan (a pretty unique program!) along with old versions of Syntrillium’s Cool Edit Pro and a bunch of other audio cleaning tools. All extremely powerful tools in their own right.

The thing is that they have all taken a fair while to learn and partially master. Along the way, someone suggested I look at Audio Cleaning Lab back at v2.0 as another application. I did and was pretty impressed at the variety of tools within it and what they can achieve fairly quickly and simply. There is a v4.0 free trial available. Maybe give that a run.

Chris Young

https://www.magix.com/au/music/sound-forge/sound-forge-audio-cleaning-lab/

MAGIX Video Sound Cleaning Lab – Tutorial - YouTube

Chris, thank you very much for the detailed comments. I will take a look at that for sure (or "fer shure" as the Californians say). Your project sounds extremely interesting! And yes, I am blown away by Capstan! True miracle software.

Today I am trying to re-laminate some laminated plastic-on-cardboard "picture discs" with hot-stamped grooves (as opposed to acetate-on-cardboard one-off actual cut discs as used in Record-O-Graph coin booths) that are coming apart at the edges. Fortunately the top plastic/adhesive layer is separating from the cardboard center, so at least I am not trying to glue the thin, flimsy plastic on its own. They are so badly spread (thickened) around the edge that they make the stylus jump out of the groove, even at 33-1/3 rpm. (I can't imagine playing them at 78 in this condition!)

And Andrew, I will post a few samples after the holiday. I need to decide which are the worst ones, so as to present a good (or bad) example and provide a good challenge.

Meanwhile, Happy Thanksgiving to those who are so inclined.

I recovered this one from a 78 rpm Parlophone a few years back and cleaned it up in Cool Edit Pro. The original Parlophone equalisation curve was applied after the RIAA equalisation curve of the reproduced recording of a modern player was removed.

Prior to RIAA standardisation of equalisation curves, there were individual flavours according to any particular recording label. Sometimes the trim of the equalisation was decided upon by the recordist's own personal judgement of what sounded best.

The stylus for a good 78rpm recovery needs to be new and for the recording type as the taper of the stylus is narrower than a LP stylus. An LP stylus will extract the recording but rides on the surface noise like pops, abrasion and old stylus slaps which tend to have jumped over the waveforms in the disk and cut new tracks closer to the surface which reproduce as distortion.

A general system seems to be to clean the record with detergent and water, taking care that no water goes into an older cardboard or woven cord substrate or it will may swell. A modern microfibre cloth is good for getting into the depth of the grooves.

Before it was discontinued, the old Mr. Sheen silicone polish spray was good. Modern Mr. Sheen is wax-based spray-on furniture polish with no abrasive content. It will still work to fill in some crackles.

The dried wax needs to be raked out with one playback pass with a worn stylus so that the sides are filled but a little material remains for the good sharp stylus to dig out of the bottom of the groove and lay into the gaps in the sides ready for the second recording pass..

After blading off the wax shreds with a paper edge, another pass brings up a little more wax which can be left alone.

A mist-coat spray with old silicone Mr. Sheen can then be applied. The new wax-based spray-on will do.

Play the record with the surface still wet with polish.

The first play through will be a dead loss as wax shreds will accrue on the stylus and deaden it. Carefully clean the stylus and the reproducer of wax.

The second play through seems to be best.

Sometimes a third play does better but after that it goes downhill.

After that it is time to do the computer work.

For the most faithful rendition, the original equalisation curves should be applied. If the recording is recovered using a modern turntable with the RIAA equalisation being applied, the RIAA equalisation will have to be removed.

However with a good set of headphones, using the graphic equaliser to personal preference in the sound may work fine without having to go to deeper troubles.

This is achieved by first using the graphic equaliser to "de-RIAA" the recording. It will sound dreadful.

Then apply the older equalisation curve with the graphic equaliser. The downside of this is that surface noise will increase along with any noise floor in the recording itself.

The RIAA curve works like Dolby NR to pre-emphasise the high frequencies so that on playback, the mostly high frequency surface noise is diminished along with the pre-emphasis when it is re-adjusted in the playback.The older equalisation curves were not as strong in reducing surface noise.

The next step is to run the various noise reduction filters, then dynamic effects like compression sparingly. Finally you may find that recording mono from the record to a stereo track will allow you to apply very short delay which sometimes has a collateral effect of reducing "apparent" surface noise.

You may have to manually map out really loud pops by expanding the wave form and applying compression along a short span to flatten pops or really digging deep and manually moving the peaks of each sample.

How long is a piece of string and how obsessive can one reasonably be without going insane after a while? A psychotic affliction was known to affect ASDIC (Sonar) operators and W/T (morse code operators) as a result of the exceptionally deep concentration required for the trained ear to extract sense from the static and enemy jamming.

The radio sufferers were sometimes referred to as "dit-happy" radio operators. The breaking point apparently occurred when operators lost confidence because they could no longer distill what was real from what they might have imagined hidden deep in the static and jamming.

THE LUTON GIRLS CHOIR YOU ARE MY HEART'S DELIGHT - YouTube

That was quite the exhaustive treatise. Obviously you've spent some time restoring old records. I have heard many of those points before, adhere to the majority of them, of necessity disregard some; a few are entirely new to me and will require a bit of digesting before being swallowed.

The records with which I'm having the most trouble are cardboard based with laminated plastic surfaces, followed by acetates, some of which also have cardboard cores. Of course acetates need to be cleaned carefully, if at all, and I won't subject the cardboard ones to any liquids. Likewise I'm unwilling to attack any of these specimens with any sort of spray (be it "lubricant," wax" or whatever). I treat the shellac and vinyl discs with a traditional wet wash and distilled water rinse, then drying w/ microfiber cloths.

Referring again to the laminated and acetate records, there's no telling what EQ or cutting stylus was used. (I had never encountered the 60º stylus angle, but my records are all US origin so I think it's safe to assume 90º.) I play them all with 3 mil conical, then undo the RIAA playback EQ. After that, I can only start with some average EQ from the period (in this case '40s and '50s) and then adjust by ear.

With all due respect, I have to disagree with your statement that
The RIAA curve works like Dolby NR ... although both share a common goal of reducing HF noise. Dolby is much more complex, being a dynamic system. In the recording chain, the HF gain is boosted only when HF levels are below a defined threshold. In the playback chain, the HF is expanded downward only when HF levels are below the threshold. This requires careful calibration between record level and playback level (Dolby Type A of course being designed for audio tape and, later, optical soundtracks). Because of the need for calibration, Dolby B found only limited adoption for audio cassettes, and was a big flop for FM broadcasting. On the other hand, RIAA, and any other phono EQ curves, have no dynamic component. The HF portion of RIAA equalization boosts the recorded HF in accordance with a predefined curve, and upon playback the HF is attenuated by the complimentary curve. In fact, the HF part of RIAA system is much like the pre- and de-emphasis used in FM radio broadcasting, and which was also used for the FM audio portion of analog TV. (Unlike RIAA, broadcast pre- and de-emphasis has no LF adjustment, the LF portion of the curve is flat.)

I don't use a graphic equalizer, rather I do all my EQ within the editing software. There are a number of click-reduction programs available, as well as multi-band FFT-based noise reduction, various conventional filters, and downward expansion used for noise reduction. I don't use the same sequence that you use for the different processes involved. This is based on my experience with the tools I'm using, which pretty well agrees with the various instruction manuals as well as with some lengthy papers about archival conservation of audio recording. I have also gone to the length of copying a note from one verse or measure of a song, and using it to replace an irreparable version of that same note in a different verse.

The process can be rewarding, but also discouraging if the record contains too many flaws that can't be eliminated.

I've been remiss about posting some samples. I'll have to dig out some samples for everyone's edification.

And while I've never gotten "dit happy" (I don't work code) I can relate to the sudden burnout and lockup of one's brain after copying and logging for six or eight straight hours of a DX contest. I have literally walked out of the shack, looked around, and couldn't recognize where I was.

Greg Miller.


You are of course entirely correct regarding noise reduction.

As for EQ, I use the EQ tool in Cool Edit Pro which was soaked up by Adobe to be Audition. Somewhere on the internet I discovered the equalisation profiles of some of the recording labels and saved the EQ settings. They are long lost in the oblivion of computers past as I have not done any sound work for some years.

And yes, some records are beyond recovery except perhaps as historic pieces warts and all, especially the old acoustic records cut before 1929?? which I think was when the electric microphone came into play.

Occasionally you may come across one which has defects at the mastering level. One such from memory was a Charlie Pride 45 which changed speed during playback. It had me pulling the transport of a jukebox to bits trying to find the fault which was not in the jukebox.

Another had a fast repetitive wow which I first thought might have been a fault in the turntable I was using. Then I moved to the notion there may have been an artificial tremolo effect and it may have been. My thinking changed to perhaps there being a drive roller or pulley have got loose on a driveshaft and running out of true. The disk itself was scratchy and I overdid the noise reduction a bit.

MARK FISHER'S ORCHESTRA. "I BRING A LOVE SONG". - YouTube

The very worst one ever was an old shellac-over-aluminium substrate disk from old drive-in cinema pre-show ads. I understand they were the 1950s equivalent of low cost consumable recording media. Once they had run their time, the owner of the drive-in used to cut them up into strips for chores like strapping wooden fences. Over time, the shellac had perished to look like crocodile skin. With this one, desperate measures were called for such as gradually tilting the turntable as the disk played so that gravity helped to prevent skips and jump-backs. At the time, Australia had a bit of a cultural cringe going and "elocution lessons" purposed towards sounding more British were a thing.

ROLEYSTONE DRIVE-IN. CHAS. E. BLANKS.wmv - YouTube

Mr. Hart:

There is an excellent (unbelievable, actually) program called Capstan which can remove wow and flutter. I've never used it myself (of course it's pricey) but I've heard demos that are quite impressive. The other option (used by at least one of the "old time radio" resellers here) is to use a turntable with no center post, and to center the discs manually to eliminate wow caused by holes that were punched slightly off-center.

As for the info about past EQ practices, two possible sources come to mind. Take a look at the documentation for Audacity, and/or the documentation for the Diamond Cut line of NR products. I'll be surprised if you don't find something relevant from at least one of those sources. [Disclaimer: I was a beta tester for Diamond Cut several years ago. I have no present connection with the company. I do feel their documentation is unusually thorough as might be helpful especially to beginners.]

Some further info about Dolby. Many years ago, someone modified a Dolby Type A card, in an attempt to make a single-ended NR playback system (downward expansion only). The amount of expansion was controlled manually by the user, rather than by some factory calibrated threshold. IIRC Dolby designated this a Cat 43 (or some such name). Also IIRC there is now a separate company writing software which they say operates more or less like the Cat 43 did. In fact I have a card cage with seven Dolby A cards, and I was tempted to modify at least some of them for manual control. But after actually trying a Cat 43 for a few weeks, I concluded that it wouldn't solve the problem I was working on at the time, so my cards remain unmodified.