HOWTO convert an MP3 to a playable, 3D printed record

Instructables user Amandaghassaei has posted a HOWTO for making a 3D printed record that plays on a regular turntable. Her method converts any digital audio file to grooves ready to print. It's a bit fuzzy, but still rather exciting! I'm waiting for the way when taking a snapshot of a vinyl disc can be the first step toward deriving its audio content, converting that back to a shapefile, and printing out a high-fidelity duplicate.

In this Instructable, I'll demonstrate how I developed a workflow that can convert any audio file, of virtually any format, into a 3D model of a record. This is far too complex a task to perform with traditional drafting-style CAD techniques, so I wrote an program to do this conversion automatically. It works by importing raw audio data, performing some calculations to generate the geometry of a record, and eventually exporting this geometry straight to the STL file format (used by all 3D printers). Most of the heavy lifting is done by Processing, an open source environment that's often used for coding interactive graphics applications. To get Processing to export to STL, I used the ModelBuilder Library written by Marius Watz (if you are into Arduino/Processing and 3D printing I highly recommend checking this out, it works great).

I've uploaded some of my complete record models to the 123D gallery as well as the Pirate Bay. Check Step 6 for a complete listing of what's there and what I plan on posting. Alternatively, you can go to Step 7 to download my code and learn how to make your own printable records from any audio file you like.

3D Printed Record


  1. The Nirvana song sounded like Kurt Cobain signing from the other side and I liked that effect. Now why did she channel hop to New Order without letting the first song finish?

    1. I can see it now, it’s going to be the wav/flac/mp3 wars all over again.  Pressed vinyl vs. printed…

  2. “I’m waiting for the way when taking a snapshot of a vinyl disc can be the first step toward deriving its audio content, converting that back to a shapefile, and printing out a high-fidelity duplicate.”

    That’s been the first step in the process since 1889.  It’s just that we haven’t gone beyond that step until recently:

    1. From that article:

      “He [Emile Berliner] shares that they’re in this particular building in Hanover, and then he recites some poetry, sings a song and counts to 20 in several languages.”

      And to think, all without Wikisource, Autotune or Google Translate! Perhaps a distant relation to Paul Bunyan?

  3. The howto forgot to mention the need for RIAA equalization.  Though you can skip this step if you disconnect your turntable from the phono preamp and plug directly into a line-in you’ll still suffer some distortion due to poor tracking.

    And to get nit-picky MP3 isn’t 16bit, it is internally 32bit float and output is set to whatever the input bitdepth was…

    EDIT: and I’m curious how much the distortion due to the printer’s (relatively) poor granularity would improve if you printed at the wider 78rpm width and used the much larger 78-standard stylus? A larger stylus should get “kicked” much less by the relatively bumpy grooves and in doing so decrease the HF distortion…

  4. Not many printers will do a 12×12 object.

    I have a completely non-electronic crank Victrola.  I’d be curious to try this to make 78’s for it.  But what printer will do objects this large?

    edit: I read the article, she has a fairly exotic 3D printer at her disposal.

  5. If you want to make your own records, why not buy a lathe?  You can probably get a used one for less than a 3D printer.

    1. Maybe with a 3D printer, one’s home is less likely to join the list of vinyl record factories that have burned down. I can’t back that up with any links (record plant fire doesn’t Google well, in my hands, at least), but I’m told it was somewhat common–particularly in places like Jamaica, with lots of small record plants–by a friend who had his own disc lathe, on which we cut some records. As he demonstrated for me, the skewings from the grooves of a single disc form a loose ball of vinyl wool, which a spark consumes alarmingly fast in a great ball of flame! Ramp up to production pace, and there’s a lot of flammable wool around, which is… on the one hand, another fun reason to use a record lathe! On the other hand… well, no, you’re right: people should get record lathes. 

      But we’re all high on 3D printing, up here in BB! Plus, all the aural artifacts in the embedded video above are the abode of weird lo-fi magic!

    1. What a shame that you took time away from feeding the hungry and curing cancer to read about it.

      1. I do feed the hungry. I feed myself.

        From an engineering standpoint, you’re spending vast resources to create entropy. You have a portable, clean recording. You expend energy and use an incredibly complex machine to create a hissy, poppy wasteful physical copy that’s less useful. You lose energy, you lose the signal to noise, you waste limited time. Maybe I just don’t get it… but it seems a waste.

        As a pre-cursor to processing the MP3, you presumably have the capacity to play it. Therefore, the record is nearly useless.

        Are you suggesting a frank discussion of technology can’t encompass how useful that technology is? I think that debate should be at the heart of the discussion of technology.

        At least a Rube Goldberg machine has aesthetic appeal. If you are going to fuck around, at least do something fun and humorous.

    2. You never know.  This could lead to someone having a “Eureka!” moment on a more weighty problem.

      Maybe not, but surely you’ve done something in your lifetime which others see as a “collosal waste of time”.  Such as sitting around and making critical comments on BoingBoing.

      1. If I sent a telegram with the content of my comment to Novogorod, had it translated into Russian, sent back to me, re-translated it into English and then finally posted it, it would be as wasteful as this process.

        It’s actually more efficient with less loss of signal to cut an acetate with the MP3 playing loudly over your computer’s speakers. MP3>Processor>Audio Output>Dictation Machine> Acetate Record vs. MP3>Processor>Digital Output>Transcription Program>Digital Output>Printer>Record

        By commenting on BoingBoing I am expressing my views in an efficient fashion (at least from a thermodynamic view).

        Even for a hobby, this technology is a waste of time. Try Legos or Ouspenskianism!

  6. This is such an obtuse way to solve this particular problem. Records are made using radial, not Cartesian coordinates. On a lathe, not a mill.

    1. Maybe, but there’s a thing called a digital painting. Brushstrokes (arcs created by motions of the fingers, wrists, and/or arm) can be laid down in arcs on a cartesian-type grid, in a raster. I’m certain the argument that it was a less than optimal solution a problem was advanced many times. But it still didn’t do much to deter its development. 

      (Or CAD plotters. Remember when they used to be printed entirely with cartridges of pens like RapidoGraphs?)
      Personally, I think this is neat as hell.

    1. I will give them it is a cool project but yeah that’s sorta how I felt when I saw the article.

      My logical/rational side says no one is dumb enough to think that this is anything more than a curiosity…. Sadly It wouldn’t surprise me if some lobbyist spoon feeds an elected official who is completely ignorant to the technology and this molehill suddenly turns into a mountain.

    1. I can only assume and hope that was sarcasm.

      Because if it wasn’t and your desk is laid out anything like mine, Look down and to your right at your CD/DVD burner. As it is far more effective/practical than 3D printing your own record will be and will likely stay that way for the next century or so.

  7. One way to drastically improve the results would be to create a radial or cylindrical 3D printer.  Basically, instead of having a platform that translates in cartesian coordiantes, it would only translate along the Z (height) axis, and then rotate around a central point, with the print head on a single axis to set the radius.  Object layers would be printed from the inside-out.  That would remove the whooshing aliasing we’re hearing here.  It would probably introduce other artifacts, though, such as a kind of aliasing when the stylus goes from an arc that is printed to an unprinted arc (as in, during printing, the platform spins, and the print head just draws “dotted arcs”, so that during playback, the stylus would end up jumping slightly every time an arc starts or ends in the area of the groove.)

  8. I have always wondered if anyone has attempted to design an optical transducer to directly produce the original audio signals from a microgroove LP. 

Comments are closed.