Industrial ASMR: Mining an Objet 3D Printer for Sound

Ever wonder what it sounds like to be inside a 3D Printer as it's printing? Well, let's find out!

First off, we'll need to pack equipment for this sound mining adventure.

tl;dr list: Sound Recorder, Microphones, Preamps. Read on if you want to experience gear nerdery.

• Sound Recorder: In this case, a Sound Devices 722 Recorder. This is the bottom of the top of the line, with 2 channels using Sound Devices fantastic mic preamps, meaning our recordings come out with only what the mic pics up. You could also use something like a Zoom H6. I really don't recommend the Zoom H4n, as while popular, it's also a noisy mess, and some of the sounds we're recording are very quiet.
• Microphones: For a machine like this, we've got 3 options for kinds of mics to use
  • Condenser Mics - In this case, I was hoping to use shotguns. These microphones have a very narrow pickup range, meaning we won't have to do too much filtering for outside noise. They can be put in the printing area to try and capture sounds of printing, though with all of the fans and motors, it may be difficult to differentiate sounds. The only shotgun I had available was an Sennheiser MKH8070, which for the small size of the print bed is just asking for trouble. Instead, I used a http://en-us.sennheiser.com/condenser-microphone-cardioid-guitar-acoustic-bass-brass-mkh-8040. This is a short cardiod, meaning we'll pick up more surrounding sounds, but it should still work pretty well.
  • Contact Mics - Contact microphones are perfect for this application. Not only do they have a small profile, they are VERY sensitive, so we can sense all sorts of tiny, interesting sounds. You can certainly make your own, and there are a few instructables on how to do so (including https://www.instructables.com/id/3-contact-mic/ and https://www.instructables.com/id/3-contact-mic/). I went with some prebuilt ones:
    

    • http://hydrophones.blogspot.com/2011/05/c-series-p...

    • http://www.aquarianaudio.com/h1a-3.html with a http://www.aquarianaudio.com/contact-mic-adapter.h... contact mic adapter

  • Induction Coil Mics - These microphones are not well known or widely used, but can be a lot of fun. Instead of picking up vibration from either air or surface, they are used for sensing electrical fields. They are usually sold as recording devices for old analog phones that used voice coils to relay sound. Making one is fairly simple, any most induction coil hooked to a pair of wires will do. However, sturdy commercial versions are available for super cheap. You can find them at places like Radio Shack for $6-10 (similar to
    

    http://www.amazon.com/Telephone-Microphone-Suction... I can't find the radio shack link right now but it might be worth checking stores). If you're in Europe and want one with some of the limiting circuitry removed, try http://hydrophones.blogspot.com/2010/09/induction-...
    

• Mic Preamps: Since we're focusing on contact mics, we'll need a way to make the mic impedance match the recorder input impedance, otherwise we'll get silence or a high pitched mess. This has been written about to death around the net, with one of my favorite version being http://www.musicofsound.co.nz/blog/the-first-rule... so I highly recommend educating yourself there if you're not familiar with the needs to contact mics.
  • If you're interested in building your own and learning a bit about preamps, there's a neat tutorial available at http://www.zachpoff.com/diy-resources/alex-rice-p...
  • If you want to buy one off the shelf, I recommend one of the following:

    • http://www.aquarianaudio.com/pa1.html

    • http://www.amazon.com/Hosa-Technology-MIT-129-Impe...

Now that we've got our equipment, where do we put it? We've basically got three options here:

• On top of the printer - For non-acoustic mics like the contact , this will work well for picking up the general vibrations of the machine. For the induction coils, it's a great place to get the sweeping sounds of the print head.
• On the print bed - This will be good for our condenser mics, and also for the contact mics to possibly get some of the lower frequencies of the motors and vibration.
• In a print - Print something with a hole in it, then stick a mic in there and see if we can hear the sound of the material coming out of the head. BECAUSE WE CAN.

So let's start with the low hanging fruit, the induction coil mic. The print head has all sorts of motors and fans in it running with different sorts of digital control, meaning we're pretty much guaranteed it will make pretty sounds for us. All we have to do is put the mic on the glass over the print bed, and boom:

But what's better than 1 mic? 2 mics above a model, spaced at about the average length between human hears! I panned these mics to reflect their position in this recording, meaning you're basically hearing what would happen if you stuck your head in the print bed and also could hear electrical fields (which you certainly wouldn't want all the time unless you like living in a world of migranes).

In summary: Induction Mics are fantastic at making the future actually sound like the future.

While I had originally wanted to use shotguns, it turns out all I had access to were REALLY long shotguns that would've been very difficult to mount around the bed of the printer. So, I just used a stereo pair of small cardiods. That said, they picked up the sound of the print head motors quite nicely, even if there was a ton of noise from the ventilation and cooling systems. In the recording below, the first half is unedited recordings. I turn noise reduction on about halfway through to try and take out some of the ventilation noise, but there's only so much you can do in an environment like this.

Before we start putting mics directly under the print head, we explored the outside and on the printer bed with them. It turned out that we got some very nice low motor tones putting the C-Series mic again the outer frame of print bed, while the H1-A with adapter picked up some nice things nearer the print head and on top of the glass.

Note that this sample doesn't have noise removal like some of the others. I'll go over this in the post processing step, but you can hear how much noise we pick up from around the print lab otherwise.

Ok, now for the grand finale: Let's see if we can get the sound of the material coming out of the print heads! My strategy for this is the model above. I measured the height of the C-Series contact mics, and made a block model that would just barely fit one of them. Then we can shove a mic in it when it's part of the way done, and listen to the rest of the print happen above it. Sounds genius, right?

Well...

Here's how the plan actually went, with the help of https://www.instructables.com/member/gabrieltaft

• Start print
• Wait until about 340 or so layers in, pause printer
• Open printer, but leave model where it is. Due to the way the objet prints, it expects a certain amount of "carpet" to be under the model, so if we just stuck a cleaned, pre-printed model in, it wouldn't be the right height.
• Clean the model while leaving it stuck to the bed. This was time consuming, fiddly work, since we had maybe 240 microns (~1/100th of an inch) worth of material above it that we didn't want to crack. Using a combination of scrapers, patience, and a vacuum, we got most of the material out.
• Only to find that the mic actually was every so slightly larger than the hole I made for it in the box, making the top layer bulge up a bit. Not ideal, but it still basically fit, so we soldiered on.
• Taped down the cable of the mic to the print bed so it wouldn't hit the print head.
• Resumed the print

Unfortunately, this ended in a LOT of clipping on the recorder, meaning something was coming directly in contact with the mic. We figured out it was the roller hitting it on each pass. The printer wasn't damaged, but it didn't sound great either. Also, there was no real way to tell if we were recording the sounds of the material coming out of the print head.

Boo.

But, when life gives you 3d printed lemons, remove them from the print bed and just put the contact mic directly where they were.

Yup, we made an even bigger mess, and it sounds like this:

You can hear the printer spin up, and then it starts breathing heavily in your ear. I believe that sound is actually the roller wheels again, except this time we've got at least 5-8mm between the mic and the wheels, so we don't get spikes from direct collisions.

Even so, there's still nothing that's obviously sounding like the material hitting the mic. I think the amount it lays down per pass is simply too small to pick up via this method. I don't think it's impossible, just more complex than I'm willing to get for the moment.

Part of the goal of this project is to try to keep the integrity of the sound, removing and changing as little as possible. That said, room noise sucks, and the 3D Print Lab is a VERY noisy room due to the air filtering systems going pretty much constantly. To remove the dull background roar from some of the contact mic samples, I use FIR filters (specifically, either ReaFIR or Izotope Rx4 Denoiser). These allow you to build a filter that "learns" what noise sounds like in a sample, and then removes it from the whole sample. It's not perfect by any means, but with some tuning it allows you to hone in on what you're looking for. Assuming you don't want to go the VST route, there's an instructable for doing noise removal in the open source swiss army knife of sound editing, Audacity: https://www.instructables.com/id/Removing-Excess-N...

Well, that's it for our 3D Printing Recording experiment. While the results may not have been exactly what we were looking for, it was still fun and we got some neat sounds out of it.

Now to find other printers to record!