Photo by aneye4apicture

Today’s idea is based on an idea I had about three years ago – about making it easier to convert your vinyl collection to digital files, so that you can listen to them on an mp3 player.

Flatbed vinyl scanner
The idea is very simple. Rather than play the record on a turntable and convert the audio to an mp3 file, you would simply place the record on a flatbed scanner, take an ultra-high resolution image of each side of the disc, and then a piece of software would analyse the grooves from the visual data, and reconstruct it as an audio file.

Given the technology, it’s easy to imagine that the software could be smart enough to distinguish between a groove and a scratch or a bit of dirt (ie: distinguish signal from noise), and make sufficiently intelligent decisions to restore the waveform accordingly.

In fact, there’s no reason you couldn’t pretty much have completely noiseless vinyl rips, with tracks separated where there’s silence between songs, and the software able to recognise and add metadata to the tracks based on its reading of the catalogue number engraved on the record by the playout groove.

There are reasons you can’t already do this:
First, scanners are the wrong size. They’re designed for documents.

Second, the image resolution and the lighting would have to be particularly good to get sufficient detail for the software to be able to interpret the shape of the engraved waveforms in the disc’s surface.

Third, the recording industry would probably find something to get cross about.

You can already do this
I mentioned this once at an academic conference, and there were some American researchers who had built something similar, but it was the size of a house, and could process about a record a week.

But theirs was to do with what they called ‘audio archeology’ – looking for clues about the past embedded in old recordings, and using electron microscopes to give data analysis of particular moments within recordings.

In fact, they didn’t seem particularly interested in listening to the music for pleasure at all.

Record-scanning for the masses
My idea is somewhere more sensible between that massive extreme – and not being able to do it at all. I genuinely believe that it’s already technologically possible to have a consumer device that can handle the high-resolution imaging that could scan records.

It might not immediately be a cheap consumer device, but for someone with a fair few records, it could be well worthwhile.

The writing of the software would be a comparatively trivial task (though I wouldn’t be able to do it).

Of course, the first thing you’d need is for the scanner to be able to take a 12″ record. The upside is that such a scanner would also be able to handle most documents as well.

Why this would be important, rather than just cool
To this day, the vast majority of records ever released have never been issued on CD. Many of them already no longer exist as playable master tapes, as magnetic tape decays over time.

But to be able to take a record and extract the sound off it without dragging a needle through the grooves, bouncing it off the bits of dirt and subjecting it to wow and flutter, you could actually analyse and restore a piece of thrift store vinyl to its studio quality – or very nearly.

And of course, you’re not restricted to albums here. 78rpm shellac discs, 7″ singles and even old flexidiscs that came with the NME or Mad Magazine would be fair game.

You make one of those bad boys, and I’ll buy one.

Table of contents for 30 Days of Ideas

  1. The other way of following first
  2. Now we’re up and dancing
  3. 30 days of ideas – 01: Keymash
  4. 30 days of ideas – 02: Radio Alerts
  5. 30 days of ideas – 03: Only Famous (a romantic comedy)
  6. 30 days of ideas – 04: Modcasts
  7. 30 days of ideas – 05: Numberless Calendar
  8. 30 days of ideas – 06: SpringCleanr
  9. 30 days of ideas – 07: Street Gallery
  10. 30 days of ideas – 08: Smart Business Cards
  11. 30 days of ideas – 09: Recordings in Concert
  12. 30 days of ideas – 10: Vinyl scanner
  13. 30 days of ideas – 11: Photo Stack-and-Scan
  14. 30 days of ideas – 12: A Box of Cool
  15. 30 days of ideas – 13: Karaoke-Tube Celebstar Idol
  16. 30 days of ideas – 14: I Made You A Tape
  17. 30 days of ideas – 15: Newspaper download codes
  18. 30 days of ideas – 16: Pebble Splash
  19. 30 days of ideas – 17: Digital radio, somewhere useful
  20. 30 days of ideas – 18: Public domain music collection
  21. 30 days of ideas – 19: Blog cast-list automator
  22. 30 days of ideas – 20: The Retirement Pile
  23. 30 days of ideas – 21: Nationalise EMI
  24. 30 days of ideas – 22: The Stainless Steel Rat (the movie)
  25. 30 days of ideas – 23: WordPress Bandcampify template
  26. 30 days of ideas – 24: Rollercoasters as public transport
  27. 30 days of ideas – 25: Next-gen personalised music radio
  28. 30 days of ideas – 26: New Music Trust
  29. 30 days of ideas – 27: Tamagotchi Gardening
  30. 30 days of ideas – 28: Charity shop clothing subscription
  31. 30 days of ideas – 29: ‘Now Playing’ social music app
  32. 30 days of ideas – 30: House of Spare Ideas
  33. Mixtape for You by Ray Kuyvenhoven
  34. What can you do in 30 days?

9 thoughts on “30 days of ideas – 10: Vinyl scanner

  1. this might also be adapted to work with a regular scanner. some sort of marker on the surface so you could scan half of it and turn it around, then knit the two images together. This is a very interesting idea :)

  2. Actually, a lot of people have been working on this problem. I first heard of it through some research done in Japan but there is even a commercial player (or maybe just a prototype) available that uses a laser:

    The sounds I’ve heard from various efforts on this are pretty lame. I actually don’t think that anyone will ever perfect the sound. The problem is that vinyl is optimized for a needle in every way — everything from the size of the grooves to the frequency range in the groove, everything. One specific problem that the Japanese researchers were running into was that they couldn’t find a way to eliminate microscopic (nano-sized) dust particles from the surface of the grooves. Needles do this by shear power: they just plow through it.

    However, I think the laser technology is promising for ancient cylinders, acetate, magazine insert records, and other historic phonographic recordings that can no longer be played with a needle.

    What I would really like to see is an easier way to digitize vinyl by way of recording it using a standard player. There are USB players for this but the sound quality sucks. There are all sorts of issues that can and should be automated: start / end of songs, metadata for songs, album artwork. All those things are necessary (to me) for a proper digital copy of a record and are time consuming to do by hand.

  3. I had the same thought as mr.voxius. Just take two scans. If the software is smart enough to decode the audio from an image, it should be able to manage knitting 2 images.

    My question is; What resolution scan would be required to reproduce the encoded audio? We have pretty good consumer scanners these days. Would 2400 dpi be sufficient?

  4. ELP does it with laser by reading it longitudinally. Competes with ultra-high end audio at double the price. We’re talking 15-20k here that competes with 30k turntables/tonearms/cartridges.
    The idea of scanning the vinyl and putting it back together is something to look at. If they could do it in realtime (and have been able for a while now), I’m sure today’s technology should manage.

  5. Considering you want to scan a groove that has intricate content for up to 20Kh (maybe more) it would ne a matter of measuring how long physically is a 20k wavelength playing at the lowest speed playable, most likely 33 1/3. Then all lower frequencies will be well covered, as well as the higher playing speeds. But this scan has to be 3D since the groove is not just a line on a vinyl. It is a V shaped cut into the vinyl that has different “bumps” on the left wall from the right… This would be the starting point of the scan that would produce a 3D picture. THEN the fun part starts: How to transform the variations of the contents of both channels without losing precious details from the original recording, while preserving phase, all of this respecting Speed, alignment and RIAA curve. I don’t say it’s impossible, far from it, but the sacanning resolution part is not the end of the problem, in my opinion. I’d like to talk more about this.

  6. I have been looking at scanning vinyl records to digital format for many years.
    I agree that the groove that the stylus follows is indeed 3D with independant side wall information corresponding to stereo left and right information.
    With this in mind it seems that close range photogrammetry techniques could be applied to sucessfully solve the problem.
    The steps would be:
    1. Capture the record surface with as many scans as practically possible with the captured resolution set at the highest optical resolution supported by the scanner. (Do not use interpolated resolution values)
    2. Use photogrammetic software to perform a camera (in this case scanner calibration) and then backdrive targets (conjugate points) visible between overlapping scan images.
    3. Build the image block network and apply bundle block adustment to arrive at a 3D data set that is statistically robust and reliable enough to exceed the finesse of measurement required to accurately capture and decode the highest audible frequencies in the record grooves.
    4. At this point one would need to model the groove surface and then write the software to produce the virtual stylus with which to decode the imprinted information within the groove.
    5. Further parameters needing to be considered are the constantly varying stylus speed as the groove is followed from start to finish along with the lead angle of the stlyus w.r.t. the groove of course.
    (These last considerations mean that the wave form of a constant pitch note becomes more and more compressed as the stylus progresses and that the stylus tip harmonics alter in addition)

    I am sure that other applications will throw up useful material to this end, such as fingerprint analysis etc.

    Hope this is of interest

  7. Hello,
    Reading all you makes me think this could be possible maybe to adapt another way to read vynil with hd video maybe.
    Imagine a cam just upon the stylus of a turn table. The cam is pointed to the vynil with focused very close with a fine optical and a prism that could catch à 45° the left and the right track. Imagine also that all the recording is time coded by the move of the turntable.

    The movie created could be decoded i think with a very much better quality becose of timecode and 3d like split screen created using the prism.

    A decoder would use this video to encode the sound.

    that would be perfect

Leave a Reply

Your email address will not be published. Required fields are marked *