Cutting records on a cnc mill

[kallaballa]

In the past I've successfully cut terrible records on the laser cutter using https://github.com/kallaballa/sndcut. Now I'd like to improve on that a bit and use my CNC mill. For that i adapted sndcut to output g-code as well and I'm planing on using a drag knife that i already have.

Now my questions:
- I can't be the first to try this... can i find any reports on this in the forum? (i tried)
- What materials could i use?
- What other methods besides a drag knife might work on a CNC mill (i tried engraving which ultimately failed)

greets,
kallaballa

[dubcutter89]

Interesting stuff - is there a sound or video of the actual cut records?
Regarding your questions:
- there's not to much going on on this site regarding new technology as most people here are doing it the oldschool (100+ years) way...
- maybe start with some plastic? I guess milling a solid hunk of gold would be rather expensive
- I could imagine that simply using a needle to emboss/impress a groove could work!

Cheers

[kallaballa]

I bet there are still some veterans that can think of advice for me... after all, this is the experimenters' & innovators' forum, isn't it?
I tried to engrave the groove on acrylic with a rotating stylus... didn't work at all.
I have the following materials at hand that i think might be worth a try: PVC, delrin, acrylic and polycarbonate.
What i would like to do next is use a very small drag knife, which is pretty similar to your idea with the needle. But if that doesn't work I'll give a needle a try

[markrob]

This is an interesting project. I assume your expectations on quality are low.

I would think a CNC mill would be far worse than a laser engraver. The advantage to using a CNC with a drag knife or embosser, is that you can get the correct groove geometry. The typical the laser engraver/cutter resulting groove geometry would seem to be uncontrollable without resorting some very specialized tricks. I would avoid using the spindle motor at all costs. Just the slop from the runout and vibration would swamp out any signal you are trying to cut. Even with the spindle not running x/y stepper motor noise could be a real issue.

I have a question about your IRIAA conversion. I assume you are treating the audio file as displacement information. Are you using the normally published IRIAA curve that shows a rising response with breakpoints at 50hz, 500hz and 2122hz? If so, this is the wrong curve to use. That curve assumes your cutter head has flat velocity response, not displacement. You must integrate the IRIAA response curve to correct it to be displacement based. Or, you could differentiate the audio to make it velocity based and then use the published curve.

I would try to fabricate a spring loaded HSS embossing tool with a conical 90 degree tip. Then you would zero the Z axis as is normally done and set the depth based on the amount of down spring force vs z axis travel needed to attain the groove width. For plastic, something in the 20-50 grams of force should give you a decent groove width. You might also have to build in a bit of backwards tilt to the tool so that it drags smoothly.

We are all waiting to see something from the HD Vinyl folks after several years of development.

Mark

[kallaballa]

Interesting stuff - is there a sound or video of the actual cut records?

I'll make a recording but it is terrible .

[kallaballa]

I assume your expectations on quality are low.

indeed

I would think a CNC mill would be far worse than a laser engraver. The advantage to using a CNC with a drag knife or embosser, is that you can get the correct groove geometry. The typical the laser engraver/cutter resulting groove geometry would seem to be uncontrollable without resorting some very specialized tricks. I would avoid using the spindle motor at all costs. Just the slop from the runout and vibration would swamp out any signal you are trying to cut. Even with the spindle not running x/y stepper motor noise could be a real issue.

My CNC mill has a higher precision than my laser cutter, so i was hoping to get a better result, somehow. Also the laser cutter vibrates a lot as well. But anyway i think the noise from the stepper motors should be inaudibly high pitched because i plan to cut with 5mm/s, shouldn't it? I don't understand what you mean by "slop from the runout".

I have a question about your IRIAA conversion. I assume you are treating the audio file as displacement information. Are you using the normally published IRIAA curve that shows a rising response with breakpoints at 50hz, 500hz and 2122hz? If so, this is the wrong curve to use. That curve assumes your cutter head has flat velocity response, not displacement. You must integrate the IRIAA response curve to correct it to be displacement based. Or, you could differentiate the audio to make it velocity based and then use the published curve.

yes, that is the curve I am using. are you saying that the frequency response of the cutting tool may differ from that of a regular vinyl cutting tool?

I would try to fabricate a spring loaded HSS embossing tool with a conical 90 degree tip. Then you would zero the Z axis as is normally done and set the depth based on the amount of down spring force vs z axis travel needed to attain the groove width. For plastic, something in the 20-50 grams of force should give you a decent groove width. You might also have to build in a bit of backwards tilt to the tool so that it drags smoothly.

interesting idea!

We are all waiting to see something from the HD Vinyl folks after several years of development.

Who are the HD Vinyl folks?

thx!

[markrob]

Hi,

Here is a link to the HD Vinyl web site.

https://hdvinyl.org/

They have invested quite a bit of time and money to do what you are describing.

I am just pointing out that that any mechanical vibrations, runout, or play in your system will get encoded into the groove along with your intended motion. Remember the maximum deflection of the groove is about +/-150um. At high frequencies, the grooves have details measured in the nm range. That's in the same range as a wavelength of light. At 5mm/s, your steppers will have torque cogging or ripple that will be in the audio range and this will very likely get encoded into the groove. I'm sure you can hear and feel the stepper pulses on your CNC as they move the gantry. If you want to get a handle on this, I would suggest cutting silent grooves (e.g. a pure spiral) to start and get a feel for the level of the noise floor. This will also help you develop the correct tool and cutting forces.

You can cut a record in a constant amplitude or constant velocity mode. This refers to the stylus lateral motion, not the rotation of the disk. For technical reasons, records are cut in different modes at different frequency ranges. At low frequencies (50 - 500hz), they are cut at constant amplitude. In the range of 500 - 2122hz, constant velocity mode is used. Above 2122Hz, the mode changes back to constant amplitude. The RIAA curve is used to make this happen. A mirror of this curve is used in the playback chain so that the result is a flat overall frequency response. The typical published curves for cutting and playback are defined assuming the cutterhead and pickup have flat frequency response vs velocity. If you are mapping the audio file data directly to groove position or deflection, then you are cutting using a constant amplitude mode. So the published curve cannot be used as defined. However, it is easy to convert the curve from velocity to amplitude mode or to process the audio file to make it velocity based. Hope that makes it more clear.

Mark

[kallaballa]

I think i understand now. I'll do my first test-cut on the weekend

[kallaballa]

Took me way longer to get around to it, but right now i am in the middle of my first attempt: https://chaos.social/@kallaballa/106702187653388281
I'll post the result as well.

[Oldskoolhousehead]

Hello, latecomer here. My question is, will laser cut records be mono or stereo? Thank you.

[thelace43]

Hi, I lurk here often.

If I understand correctly, the discussion is about using a CNC to cut a grove based on an audio file.

I have been thinking about something similar, but it would bypass the CNC and just cut a record with a laser instead of a stylus.

I don't know much about lasers and CNC, but since it's been on my mind for the last few months (I'm designing some lathes and a couple cutting heads) I thought I'd chime in.

[CutR]

I have been thinking about something similar, but it would bypass the CNC and just cut a record with a laser instead of a stylus.

This guy tried what you looking to do http://www.williamosman.com/2017/03/laser-engraving-records.html

[thelace43]

[/quote]
This guy tried what you looking to do http://www.williamosman.com/2017/03/laser-engraving-records.html
[/quote]

Thanks! I wasn't thinking about a CNC laser, but rather one that could focus to a smaller point.. After some recent thought, and knowing what real grooves look like, I don't think it would really work at all, unless it could be read back by a laser and converted into a digital signal somehow.. ; )

EDIT: oops, I guess I don't need to delete anything when I quote-respond

[Oldskoolhousehead]

Hello. My question is, which Linux package should I use on Monterey OS? Thanks in advance.

[sdt9030]

I have a Snapmaker 2 multifunction 3D Printer/CNC/Laser Engraver and have now used it successfully to cut records playable on my 1924 Victrola using 3 techniques: the CNC approach, a spring-loaded diamond engraver (in place of the CNC cutting tool), and more recently the laser engraver to cut grooves into different plastics (PVC, acrylic, acetal). Each approach has its pluses and minuses, but I am leaning strongly toward the laser approach as having the best ultimate potential. The SN2 doesn't have the spatial resolution to get the same groove density as in a LP, but it is good enough to probably get to about a 3 minute recording time on a 12" 78rpm record.

SteveT

[sdt9030]

Quick followup SAFETY note to my post above. DO NOT USE A LASER to engrave or cut grooves in PVC! I should have realized that the PVC would give off chlorine gas when hit with the laser but I didn't think it through. Fortunately I stopped soon enough to avoid damaging either my laser system or my lungs! Anyway, I'm still using the laser, but I am now restricting myself to materials that are safe for laser cutting/engraving.

SteveT

[kallaballa]

I've written a program that generates laser cutter models of LP records that you can cut, if you are still interested

https://github.com/kallaballa/sndcut

[alexdow]

free Perfect Groove v2.1.0
As you might have read, the HD Vinyl project is currently on hold. Since Perfect Groove was born as part of that project, we sadly had to halt development of the software as well.

It was a fun time working on such an outstanding and challenging project and we had the opportunity to meet and collaborate with a lot of top-notch cutting engineers, mastering engineers and other specialists. Thank you all for your support and insights!

We have decided to release the latest version of Perfect Groove free of charge to anyone interested in the process of vinyl cutting. You can download your copy of Perfect Groove v2.1.0 and use the credentials below to activate a perpetual license.

Happy cutting and see you soon! https://www.perfectgroove.com/

4-nanometer precision and 50-60 nanometers!!!

Austrian music company founder Guenter Loibl has been forced to suspend his ambitious ‘HD Vinyl’ research initiative due to disappointing audio fidelity tests, according to the company and filings with the Austrian government.
The HD Vinyl project was first hatched as a separately-funded R&D initiative within Rebeat Innovation GmBH in 2018, with Rebeat founder and CEO Guenter Loibl personally investing the most substantial tranche of funds into the project.

Rebeat, which started as a physical recording distributor in the early 2000s, has since expanded into a diversified portfolio of music technology services and companies. That includes a suite of digital distribution services, royalty accounting solutions, a company focused on verifying data from streaming platforms, and development initiatives like HD Vinyl. The company remains a stronghold in the Eastern European music industry, with substantial expansions into broader Europe, the United States, and portions of Asia.

The vision of HD Vinyl was to dramatically improve the audio quality of conventional vinyl records by using a sophisticated laser-etching process. Initially, the more granular and fine-tuned grooves promised to deliver extremely high-fidelity audio without sacrificing the warmth and cool-factor of vinyl LPs. Even existing turntables would be compatible with HD Vinyl releases. But despite the theoretical possibilities, Loibl explained that current digital laser technology proved incapable of delivering the precision and dynamic range that listeners would demand.

“A lacquer has a dynamic range of up to 90db,” Loibl told Digital Music News. “The most that playback systems on turntables can reproduce is around 60-65db. We achieved 50-55db, which simply isn’t enough for a high-quality audio product.”

By Austrian law, Rebeat has been forced to file its HD Vinyl initiative into administration given the number of investors and outstanding Austrian government-supplied loans for the project. Loibl explained that proper payback of Austrian government-funding would be impossible within the six month timeframe given the testing concerns and resulting time-to-market delays.

Loibl also noted that a pivot is now underway while HD Vinyl is on pause, with the underlying technology and approach being thoroughly reexamined. “HD Vinyl is definitely not dead,” Loibl assured, though he noted that the existing approach proved insufficient to achieve the ambitious audio-fidelity goals envisioned.

So what went wrong? According to Loibl, HD Vinyl stumbled upon a surprising discovery: analog cutting technologies are tough to beat, even with the most sophisticated, modern-day technologies.
“Neumann was a genius back in the day when he built his cutting lathe,” Loibl elaborated. “This 60-plus year old machine produces unmatched precision when cutting lacquers.”

“Just one example: a well-calibrated and well-maintained cutting lathe can cut modulations smaller than 1 nanometer. You cannot see that modulation in a microscope anymore, but you can still hear it. Just for comparison, the most sophisticated microchip production process currently delivers 4-nanometer precision, which is 5-10 times less precise than a cutting lathe.”

The sobering discovery unfortunately took a while to realize. “Using all the magic possible, we could only achieve a precision of 50-60 nanometers — and this is only in the best case scenario,” Loibl noted. “That means that a digitally-controlled laser system cannot overtake an analog cutting lathe when it comes to sound quality. I guess some vinyl aficionados already knew that regarding sound quality — but the tools to achieve such a precision digitally will not be available for the next 15-20 years. So we will need a different technical approach for HD Vinyl.”

Loibl indicated that planning is already underway for a technological reboot involving a more analog approach. ”Later this year, we’ll do some testing with an analog-controlled laser,” Loibl said. “If these tests are successful, we will restart the HD Vinyl project.”

Meanwhile, vinyl record sales remain robust and growing in many parts of the world.
According to data from Luminate, vinyl record sales topped 19.4 million units during the first half of 2022 in the United States alone, up slightly from the same period in 2021. That volume is still small compared to streaming revenues, though it’s enough to support a revived micro-economy within the music industry.

Alongside his efforts developing HD Vinyl, Loibl is also president and one of the founders of the Vinyl Alliance, a trade group that has materialized around the resurgent vinyl record industry with members like Universal Music Group, Sony Music Entertainment, Warner Music Group, Ortofon, Pro-Ject, Audio-Technica, GZ Vinyl, Record Store Day, VinylMePlease, Gotta Groove Records, and others.

“Sometimes failure is necessary, because it teaches us what doesn’t work,” Loibl mused. “But with the huge amount of knowledge we’ve gained about vinyl production, we have a shortcut on future development.”

https://www.digitalmusicnews.com/2022/08/19/hd-vinyl-rebeat-pause/

[farmersplow]

free Perfect Groove v2.1.0
As you might have read, the HD Vinyl project is currently on hold. Since Perfect Groove was born as part of that project, we sadly had to halt development of the software as well.

It was a fun time working on such an outstanding and challenging project and we had the opportunity to meet and collaborate with a lot of top-notch cutting engineers, mastering engineers and other specialists. Thank you all for your support and insights! .....

In my view, the most innovative project in the history of recordings in the last 50 years! I think with the acquired knowledge the project will eventually work by means of an analogue-controlled laser... even if it will still take a few years.

[OptVox]

“Just one example: a well-calibrated and well-maintained cutting lathe can cut modulations smaller than 1 nanometer. You cannot see that modulation in a microscope anymore, but you can still hear it. Just for comparison, the most sophisticated microchip production process currently delivers 4-nanometer precision, which is 5-10 times less precise than a cutting lathe.”



https://www.digitalmusicnews.com/2022/08/19/hd-vinyl-rebeat-pause/

I think that the 1 nanometer mark as a reference of anything most be re-explained. A one-nanometer mark fits 550,000 times in a second (ultra-ultra-sound...), and its level would be inaudible.

That 1 nm mark can be anything... If a diamond cutting stylus has a diameter of 10 micrometres , that means that the 1 nanometer mark is maybe caused by the friction of the material with the stylus in general, or something else... but that's the imperfection. The actual signal over noise is going to use a lot more area and depth, and be more defined.

It is very different to rotate a disk and make a mechanical recording in real time than print it like a pixel image.