Stero channels aren't encoded in the depth and radial axes. They're encoded in the same plane, but rotated 45 degrees, so a mono record player (depth only) plays a balanced sum of each channel.
When I read the title I immediately thought of a "laser tip" that and reading the reflections of it, not something based on software that would "read" an an image.
Following this animation I'm left thinking that you could do a bit of filtering to better extract the wave(s) from an image, though you'd require a particular setup for the lighting conditions to get the best possible quality out of it. And then back to my first idea: probably a laser-tip would be the best tool here too.
Unfortunately I don't have vinyl records around to test any of this!
Old Edison disc records encoded single-channel audio in the vertical axis (varying the depth), but regular 78s and single-channel vinyl records encode the audio in the horizontal axis.
Carl Haber, a guy I worked with at LBNL[1] about twenty years ago pioneered this technique, at least as far as I know. The really cool thing about it is how it can be used to play very fragile recordings without damaging them.
"The method has been used to successfully play several recordings for the first time, including an 1860 phonautogram, the oldest known sound recording of a human voice, and the only known recording of Alexander Graham Bell’s voice."
These are so wonderfully distorted. I am getting somewhere between the signal in the movie "Contact" and whispers of the hive mind of the Borg. Amazing sounds, even if they aren't true to the recording.
Aside from the slight wah wah effect they are pretty great. I immediately loaded them up in my S2400 and resampled them to 12bit 26khz to add much aliasing Sample C2 is especially haunting this way.
I actually got to hear one once. Extremely neutral and high-res, but the record has to be super clean because with the laser, you will hear every little speck of dust. Also it doesn't work with colored vinyl, iirc.
edit: they use 5 different lasers to get the data out of the groove, go read the section on their site, it's actually really intricate and cool
> the format has some funky EQ and other things to make a physical needle work better
Yes, in order to fit the high bandwidth music signal into a groove with a fixed width, the bass is EQed down and treble EQed up, since for a constant velocity sine wave, a low frequency wave would take up more physical width on the record than a high frequency wave. Without this EQ, you would only be able to store a small amount of music on a record. When the record is played, your phono preamp applies an inverted EQ curve to restore the signal to its original frequency response.
In the past, there were lots of different vinyl equalization curves, but the RIAA curve has been standard for decades now.
The problem is dirt, which sounds like a minor issue, but it renders the concept almost completely useless.
A stylus pushes dirt out of the groove as the record turns. A laser cannot do so, and it's virtually impossible to fully clean a record.
The only remaining advantage is that a laser turntable does not wear out the disc, and it's an incredibly minor advantage, compared to simply playing a record once on a traditional player, and recording the output.
Slide/negative scanners use multiple passes and depth perception to cleanly ignore dust on film. I'm surprised that this wasn't part of ELP's spec to the engineers. A giant fail if they can't even remove this noise in post-processing.
I wonder if they could do what film scanners do - they use an IR channel.
EDIT: and as soon as I submitted I realize film scanners scan for light and maybe use the IR channel for physical scanning. with a record everything is physical.
or maybe the IR channel focuses on a different plane? could they use different wavelengths for distance sensing/separation?
The funky EQ is the RIAA curve[1] - by essentially rolling off the bass and boosting the treble you can fit more on a disc, it attenuates some of the clicks and pops and is kinder on the vinyl and needle.
It's trivial to invert in software, and as another response points out, there is indeed a working laser record player - although the record needs to be extensively cleaned before playback or it doesn't work very well.
Thinking about a simpler system... how about getting a camera on a turntable recording a video while the record turns. It would then be possible to track a line with an specified width in the video and turn the changes in a piece of this line it into a waveform.
Also capture the audio sample from the same record alongside the video. Repeat with lots of records. Train a model to predict the audio from the video. Voila! Play records with just the camera “stylus”...
Random Q: do I have a damaged/bent stylus if a certain record (Toots and the Maytals' Pressure Drop picture disc) goes silent at several different times? I checked the stylus force scale at the different test points, and it's right in the middle of the cartridge specs.
Every other record plays fine, except some new records play slow and/or some carve out a thin hair of lacquer/enamel the first time. This always makes me think some sort of damage is happening from a misaligned cartridge. But, other good quality and used records don't do this and work fine.
If it is just happening with a specific record (especially a picture disc) it's likely not your stylus and just the pressing. You can give cleaning your record and stylus a shot and see if that helps.
If records are playing slow make sure if your table is belt driven your belts are still in good condition. Also make sure your tonearm height and anti-skate are properly adjusted.
Cleaning your records before you play them (even brand new records) is important.
It's a direct-drive Panasonic PL-5 with the internal calibration strobe, so no belt.
I always clean them with a Zerostat and Discwasher. And have a Japanese stylus cleaner too (the stuff that looks like a breast implant for mice porn stars.)
The local record store has the same record, maybe they'd compare them?
Now I'm dreaming about a program that would simply play a high-res photograph (scan) of a vinyl disk...
This also reminds me of the idea to use paper for archival. This is based on two factors: that modern office paper is extremely durable (and is likely to outlast most of the other physical media out there), and that scanners are (and have been for years now) extremely high-resolution; so, you can simply print the bits (say, of a zip file) as dots, and then you could use a scanner to read the file back!
Interesting concept the biggest problem would probably be data density. Like how many papers would you need to to get about dvd size. Other than than the problems in getting a reader dvds is one of the best current long time solution.
Readit News
Edit: mono is radial only, not depth.
https://www.vinylrecorder.com/stereo.html
Following this animation I'm left thinking that you could do a bit of filtering to better extract the wave(s) from an image, though you'd require a particular setup for the lighting conditions to get the best possible quality out of it. And then back to my first idea: probably a laser-tip would be the best tool here too.
Unfortunately I don't have vinyl records around to test any of this!
You mean horizontal only?
Old Edison disc records encoded single-channel audio in the vertical axis (varying the depth), but regular 78s and single-channel vinyl records encode the audio in the horizontal axis.
"The method has been used to successfully play several recordings for the first time, including an 1860 phonautogram, the oldest known sound recording of a human voice, and the only known recording of Alexander Graham Bell’s voice."
[1] https://www.macfound.org/fellows/class-of-2013/carl-haber
https://news.ycombinator.com/item?id=26959024
https://en.wikipedia.org/wiki/The_Caretaker_(musician)
https://www.elpj.com/
I actually got to hear one once. Extremely neutral and high-res, but the record has to be super clean because with the laser, you will hear every little speck of dust. Also it doesn't work with colored vinyl, iirc.
edit: they use 5 different lasers to get the data out of the groove, go read the section on their site, it's actually really intricate and cool
> the format has some funky EQ and other things to make a physical needle work better
Yes, in order to fit the high bandwidth music signal into a groove with a fixed width, the bass is EQed down and treble EQed up, since for a constant velocity sine wave, a low frequency wave would take up more physical width on the record than a high frequency wave. Without this EQ, you would only be able to store a small amount of music on a record. When the record is played, your phono preamp applies an inverted EQ curve to restore the signal to its original frequency response.
In the past, there were lots of different vinyl equalization curves, but the RIAA curve has been standard for decades now.
https://en.wikipedia.org/wiki/RIAA_equalization
Audiophiles likely know of this, but I'm still surprised.
Clean vinyl records with glue...
https://everyrecordtellsastory.com/2015/05/25/how-to-deep-cl...
It does not work great.
The problem is dirt, which sounds like a minor issue, but it renders the concept almost completely useless.
A stylus pushes dirt out of the groove as the record turns. A laser cannot do so, and it's virtually impossible to fully clean a record.
The only remaining advantage is that a laser turntable does not wear out the disc, and it's an incredibly minor advantage, compared to simply playing a record once on a traditional player, and recording the output.
EDIT: and as soon as I submitted I realize film scanners scan for light and maybe use the IR channel for physical scanning. with a record everything is physical.
or maybe the IR channel focuses on a different plane? could they use different wavelengths for distance sensing/separation?
Almost all the noise on a record is high frequency...so by applying a strong treble rolloff most of it just disappears.
It's trivial to invert in software, and as another response points out, there is indeed a working laser record player - although the record needs to be extensively cleaned before playback or it doesn't work very well.
[1] https://en.wikipedia.org/wiki/RIAA_equalization
The data has to be processed very quickly. I believe you should think in terms of 0.3m groove per second.
If you put a camera above the groove you also need something like a stepper motor to control the arm.
But I believe the biggest problem is dust. A needle will move small particles out of the way, but a camera does not.
This is also why laser turntables need 'clipping' filters. Dust will create all kinds of noise you don't want to hear.
How about designing an arm that used more than one laser 'needle'?
Every other record plays fine, except some new records play slow and/or some carve out a thin hair of lacquer/enamel the first time. This always makes me think some sort of damage is happening from a misaligned cartridge. But, other good quality and used records don't do this and work fine.
It's an AT microline cartridge.
If records are playing slow make sure if your table is belt driven your belts are still in good condition. Also make sure your tonearm height and anti-skate are properly adjusted.
Cleaning your records before you play them (even brand new records) is important.
I always clean them with a Zerostat and Discwasher. And have a Japanese stylus cleaner too (the stuff that looks like a breast implant for mice porn stars.)
The local record store has the same record, maybe they'd compare them?
This also reminds me of the idea to use paper for archival. This is based on two factors: that modern office paper is extremely durable (and is likely to outlast most of the other physical media out there), and that scanners are (and have been for years now) extremely high-resolution; so, you can simply print the bits (say, of a zip file) as dots, and then you could use a scanner to read the file back!
https://archive.org/details/stereo-review-presents-stereo-bu...