Techmoan video: Shaving Compact Discs to improve the sound

[Ed.Kloonk]

Today I'm using a special lathe that cuts a 36º edge into a CD to improve the sound quality. Is this madness...I aim to find out.

[pcprogrammer]

Yes that is one for the audiophoolary to enjoy.

[golden_labels]

You are ignorant of physics. His Audacity experiment shows no difference only because computers represent waveforms as numbers. The program was operating on a squareish approximation of a true analog wave, rounding values to the nearest integers. Not to mention the recording device is using cheap ADC, that exaggerates that error. If you compared both outputs with a precise analog instrument, like your ears, you would clearly hear the difference. /S

[pcprogrammer]

There was no ADC, because he just used the optical fiber output, so all digital. A simpler test he could have done was to stick the CD's in a computer CD reader and get the .wav files directly from the CD and use a binary compare of the two files. No need to fiddle with Audacity.

No doubt in my mind they will be identical, physics or not. And like he said in the video, when a CD has read errors it can't recover it will cause hiccups not less bass or treble.

And yes, most likely you are pulling my leg

[TimFox]

Around 1990, besides the great Armor All^TM on CDs craze, there was also the Green Magic Marker on CD rim procedure.
From a fact-check site:  https://www.snopes.com/fact-check/bewaring-of-the-green/

[Haenk]

Not only is this (of course) nonsense, but it will in fact damage your CD over time. The metallic layer is sealed, removing the sealant might/will lead to disc rot.

And: If this would make a difference on your CP player, it certainly requires servicing - it's broken 

[ataradov]

Too much effort. My $900 magic stones placed around the room do the same thing.

To be fair, we don't really know what are the real profits in that industry. But then again, people buy JPEGs of monkeys for a lot more money.

[AVGresponding]

Around 1990, besides the great Armor All^TM on CDs craze, there was also the Green Magic Marker on CD rim procedure.
From a fact-check site:  https://www.snopes.com/fact-check/bewaring-of-the-green/

The marker pen bollocks was even touted around in respectable Hi-Fi publications, back in the day (no I never tried it).

[golden_labels]

You know, what makes me simultaneously dying from laugher and pitying some of people, who meticulously shaved their CDs and painted the edges? Physical albums in good condition are having financial value. Imagine there is a persone out there, who has a bunch of now rare CDs, which would be wort 5–10× the original price. If not that they shaved the discs.

Producer’s website also claims the technology improves data media. I will not ask if it increases resolution of my JPEGs. Instead I would ask how a disc, shaved with precision worse than offered by a medieval lathe, behaves at 52× speed. I will not try testing that in my drive.

[TimFox]

American coins (dimes and above) are still "milled" or "reeded" around the edges, a procedure dating back to when one could shave a bit of silver from each of many coins to amass silver metal at the expense of honest citizens.

[helio0centra@gmail.com]

With a digital format you either get it or you don't. There's no varying quality like analog. If your CD or laser is bad it skips.

[fourfathom]

With a digital format you either get it or you don't. There's no varying quality like analog. If your CD or laser is bad it skips.

Not exactly.  CDs use Reed-Solomon FEC, so most bit errors get fixed losslessly.  But if the errors are too much for the FEC then the next recourse is interpolation through the bad patch.  If the damage is too bad then you get a digital skip.  Interpolation is lossy and might possibly be noticeable to a "golden ear" (not mine).  So, if there were some way to reduce the need for interpolation then the audio quality might be improved.

Don't get me wrong:  green markers, beveled edges, weighted CD platters, and all that other crap do nothing but provide a way to spend more money on your system.  But CDs aren't all or nothing, there's a zone there.  Probably imperceptible, but it's there.

[pcprogrammer]

I wonder about that interpolation bit if that also applies for a CD reader in your computer when it reads "digital data" and not "audio data". I assume that it will only do this interpolation when it has read part of the audio data stream, detects that there are errors, tries to resolve the errors, can't and will then interpolate between previous and next samples.

Which means that it is software, and that it can depend on commands send to the drive, and when reading data as a binary file it will not use the interpolation scheme.

Furthermore a brand new CD will most likely be read almost error free and renders Techmoans tests somewhat invalid. The difficulty in doing a test that can be repeated in a valid manner, is that creating two exactly the same damaged CD's to invoke the same error correction is almost impossible. With a single CD that has errors it can only be done once, and it might improve back to the original behavior, but there is no way of proving that it is actually due to the shaving.

Interesting to think about though.

[fourfathom]

I wonder about that interpolation bit if that also applies for a CD reader in your computer when it reads "digital data" and not "audio data". I assume that it will only do this interpolation when it has read part of the audio data stream, detects that there are errors, tries to resolve the errors, can't and will then interpolate between previous and next samples.

Yes, that's how I understand it.  I should have said "audio CDs".  I honestly don't know what type of error correction data CDs use, but there is a spec for it (Yellow Book?)  I can't imagine that they would do interpolation on data.

[Jr460]

Hang on a second.

All digital, nothing changes, it is all the same.   Correct, but......

Music, making, it playing it, mixing it live for people, etc is a side line.

We have a little problem with jitter.   Now this came up because of AES/SPDIF digital formats.   They have no fixed sampling rates.  The sampleword/bit clock is recovered from the data stream.   Based on the data the sample clock can shift a bitt back and forth, jitter.   Some people say they an hear it all the time.   It is regarded as problem in pro audio that pro devices have an BNC jack to two to accept to to provide word clock to other devices.

Jitter in AES/SPDIF is very subtle.   it is a slight FM modulation of all the special components.   The same thins happens with tape to a turn table, but is listed as wow and fultter.

Jitter in digital audio systems is real, but the interface chips for along time have done things to reduce it, as talked abut already an external word clock, or using  a ddifern PLL to cleanup word clock from the bit recovery clocking.

No is where we get to the snake oil.   A Cd spins at it's rate can vary, thus the digital signal coming from the CD/DVD will have jitter.   That is true, however after all the error correction the bit clock used to recover the data is not the same or related to the clock that pumps out the samples.   For a CD that is defined as 44.1K.   If the CD speeds up, slows down, nothing changes the rate the samples are shipped to the D/A convertor.   This means no jitter from mechanical system.


Cut, trim, green maker all you want, either the optics pick up the pits and it turns into a sample, or it doesn't.

[ataradov]

If the CD speeds up, slows down, nothing changes the rate the samples are shipped to the D/A convertor.

And the samples are "shipped" based on some magical jitter-free clock source?

That 44.1 kHz generated from the local clock source, it is not informed by anything on the CD. So the quality of the clock going to SPDIF and DAC is the same.  SPDIF would incur some clock recovery jitter, of course, but it would not necessarily be worse than the source jitter. There is a good chance that receiver would be able to celan up the clock and would produce cleaner signal than DAC from the CD directly.

I'd also like to see credible sources showing that jitter matters in any way.

[pcprogrammer]

I guess the digital jitter will create some harmonics but with modern oversampling and digital filter techniques they will hardly be audible.

And don't forget it is all based on individual perception anyways. A recording made of some orchestra will almost always differ form the real thing. There are so many factors in the recording chain. You have the microphones, pre-amplifiers, mixing desk, equalizers, master recorder and most important the sound engineer. If the sound engineer is not so good, the recording will suck.

The fun part in it all is that the audiophiles often state that the warmth is in the imperfections. Take the debate about analog versus digital synthesizers. "The analog ones sound so much warmer" is the claim. And why, because of the imperfections. Drift between the oscillators, differences in the filter characteristics between the channels, etc. A good digital imitation takes this into account

[golden_labels]

American coins (dimes and above) are still "milled" or "reeded" around the edges, a procedure dating back to when one could shave a bit of silver from each of many coins to amass silver metal at the expense of honest citizens.

Only if the shaving was so extreme that it would put the recipient at risk of having the coin rejected by the next person/institution downstream.

Gold and silver are not very durable and coins made of them are losing material or get damaged in circulation due to normal wear. They are so prone to damage, that one method to shave coins was by shaking violently a bag with money and collecting dust produced when coins hit each other. So it was normal that most currency was missing some material and no one would care, except for the aforementioned extreme cases. It became even less relevant when coins started to be used for their nominal value and not actual precious metal content.

The violation was not against other people, but against the treasury. People engaging in shaving were doing something, that was perceived as stealing from the state, or acquiring wealth without paying taxes. And you know there is no greater crime on this planet than not paying taxes.

[Haenk]

I'd also like to see credible sources showing that jitter matters in any way.

Sorry, no link to quote, but I read a while back, that this is in fact a huge problem in production. Solution is to use a master clock and sync all digital inputs to it.

[SiliconWizard]

The whole idea if you want *objectively* better than audio CDs without resorting to voodoo (because 1. they are still only 16-bit, 2. as said above, yes they often have read errors depending on the state of the CD, how much dirt and scratching there is, how dirty the laser lens is as well, etc, in which case the CD player will correct errors to an extent and this may degrade audio quality obviously when then BER is greater than what can be fixed losslessly, and 3. yes you can have nasty jitter issues if the player is meh)...

just don't use CDs. Just play 24-bit audio using any basic computer (could be a RPi or whatever) through a decent DAC, and you'll be eons above anything a $100k+ voodoo CD setup can buy you for a couple hundreds bucks if that.

[fourfathom]

I'd also like to see credible sources showing that jitter matters in any way.

Sorry, no link to quote, but I read a while back, that this is in fact a huge problem in production. Solution is to use a master clock and sync all digital inputs to it.

This has always puzzled me.  I have a friend who is an occasional recording engineer, and he was always claiming that reading CD audio off a regular non-synched drive caused big problems.  (Of course he also used green marking pens back in the day).  As I see it, as long as the CD drive can read faster than the stream rate then the data is read in buffered chunks and no fancy clocking is required.  If the format is audio CD then the implied sampling rate is 44.1 KHz and that's all we need to know. Once processing is done, if you want to drive an ADC to get an audio signal then the only place jitter can be introduced is the ADC clock.

Now I can conceive of a system designed so badly that precision clocks are required at each interface, but is that actually the case?

[pcprogrammer]

The whole idea if you want *objectively* better than audio CDs without resorting to voodoo (because 1. they are still only 16-bit, 2. as said above, yes they often have read errors depending on the state of the CD, how much dirt and scratching there is, how dirty the laser lens is as well, etc, in which case the CD player will correct errors to an extent and this may degrade audio quality obviously when then BER is greater than what can be fixed losslessly, and 3. yes you can have nasty jitter issues if the player is meh)...

just don't use CDs. Just play 24-bit audio using any basic computer (could be a RPi or whatever) through a decent DAC, and you'll be eons above anything a $100k+ voodoo CD setup can buy you for a couple hundreds bucks if that.

As nice as that sounds, you still have to source your audio you want to listen to. And as long as that is 16bit 44.1KSa/s up scaling to 24bits won't make it better. But it seemed most of us did not care much about that quality because we turned to MP3 players and MP3 is a compression with loss of quality. Depends a bit on the selected rate but loss it gives.

Then came flac and ape, which are lossless algorithms but still based on 16bit 44.1KSa/s for as far as I'm aware.

So you would need "hi definition" audio with 24bit and maybe higher sample rates. However the last CD I bought was ages ago and I transferred all the ones I have to my hard disk and that is what I listen to when I want music. No idea what the latest state of technology is in respect to audio recordings.

I'm no audiophile and will never buy expensive amplifiers, special cables or hefty speakers, because I most likely won't hear the difference anyway. I play my music on my computer through a Cambridge Computer sound system, with one small bass speaker and 4 little satellite speakers, placed on the ground obscured by couches. And for me it is perfect as background sound.

[pcprogrammer]

I'd also like to see credible sources showing that jitter matters in any way.

Sorry, no link to quote, but I read a while back, that this is in fact a huge problem in production. Solution is to use a master clock and sync all digital inputs to it.

This has always puzzled me.  I have a friend who is an occasional recording engineer, and he was always claiming that reading CD audio off a regular non-synched drive caused big problems.  (Of course he also used green marking pens back in the day).  As I see it, as long as the CD drive can read faster than the stream rate then the data is read in buffered chunks and no fancy clocking is required.  If the format is audio CD then the implied sampling rate is 44.1 KHz and that's all we need to know. Once processing is done, if you want to drive an ADC to get an audio signal then the only place jitter can be introduced is the ADC clock.

Now I can conceive of a system designed so badly that precision clocks are required at each interface, but is that actually the case?

I think this has to do with digital mixing systems where the streams are all digital from different sources. Then it might need a master clock to make it work. With just a single digital source connected to a digital amplifier, I guess the data stream will just be buffered a bit (maybe 1ms or so) and send to the DAC at regular intervals based on the sample clock. As long as the data comes in at a steady pace a little bit of jitter does not matter.

[fourfathom]

Now I can conceive of a system designed so badly that precision clocks are required at each interface, but is that actually the case?

I think this has to do with digital mixing systems where the streams are all digital from different sources. Then it might need a master clock to make it work.

OK, I can see that.  If you use buffering for these non-master-clocked interfaces then you need to embed a timestamp or timing feedback, and I guess the format doesn't provide that.  And in a multi-source system with no master clock you will eventually get slips or overruns.  Reminds me of the telecom networks I used to design equipment for.  We did all sorts of mapping tricks to carry quasi-synchronous payloads through synchronous networks.

[pcprogrammer]

I think that nowadays it will all be done on a computer where all the data is present on the hard disks and tracks are shifted in time and cut in pieces and then mixed together into a final track.

There are loads of youtube videos of new music producers that use software like Ableton live, where they record the different tracks and then produce a final song. Like for instance Rachel K Collier. https://www.youtube.com/c/RachelKCollierRKC