80KV’s Guide to Building Wearables that Aren’t Stupid

I’ve been working with custom-developed wearable tech for about three years now, and it seems to be the aspect of my performance that audiences are most excited about. Well-executed wearables are pretty fascinating things, and I’ve been using them long enough that I’ve developed a lot of opinions about the best ways to use them. So, I present to you:

80KV’s Guide to Building Wearables that (Hopefully) Aren’t Stupid

Since we’re living in the future, there are a lot of consumer wearables available to buy. It’s certainly easier to go that route. They have advantages like being designed to be user-friendly (it’ll be a lot closer to plug-n-play than anything you build custom), they’re likely more reliable (depending on your level of confidence in your EE/CS skills), and often supported by customer service teams. They have the disadvantages of being way, way more expensive, of locking you into a certain set of physical actions that they detect, and of, well, being consumer devices; lots of other people will have the exact same thing you’re using, you won’t get to decide how it looks, and so on.

My suggestions will focus on custom devices, because that’s what I do and because I think it’s more interesting, but people working with prefab devices may also find some of this useful.

Start with a goal

I was taking guitar lessons at the same time that I was building the first iteration of the gloves, so the initial idea was that I wanted to capture the motions I was naturally making while playing the guitar — strumming, the finger patterns of chords — and turn that into accompaniment. In retrospect, I’m very, very glad I took that approach. It limited the scope, gave me a definitive end goal at which point the device would be finished, and gave me a specific set of things to research: how to detect bent fingers on my left hand, motion on my right, and convert that into sound.

Less is more

Did you ever see photos or videos of old-timey one man bands? The ones where someone is holding an accordion or something, singing, blowing a harmonica, stomping their feet to hit a drum, and so on?

Pretty cool, huh?

It absolutely isn’t. That is the goofiest shit in the world. Do not go that route. There are two good reasons to limit the range of motions that your wearables can detect: first, because it makes it easier to remember what any individual motion or combination of motions will do, and second, because wearables on all your limbs starts to look pretty stupid pretty fast unless you’re a professional dancer with incredible coordination and grace (or Onyx Ashanti, who somehow pulls it off. I think he is a wizard).

Detecting a small number of simple actions will still (thanks to the magic of combinatorics) provide you with plenty of controls to use. It will also let you remember what each gesture does, and focus on building up muscle memory so that you can repeat the gestures at will to the same effect.

Embrace the emergent properties

When you write software, you almost never understand exactly how it’s going to work, and that goes double when you’re building something that can be interacted with in any number of ways. As you use the devices more and more, you’ll discover things that you didn’t explicitly plan for — bugs, essentially — and sometimes those provide you with extra functionality that you never would have thought of.

In the second version of the Scalar Glove, I introduced a calibration routine to detect what the base resistance of the flex sensor was, rather than relying on a hard-coded number. A result of this (and the fact that I was taking the absolute value of the difference between the baseline and current measurement) was that if I ran the calibration routine while one of my fingers was bent, I could effectively invert the range — make the sensor read as on while the finger was straight, and off when bent. I hadn’t planned for that, but it expanded the range of the ways that I could use the gloves, so I kept it.

Notice how you use them, then expand the language accordingly

As you use your wearables, pay attention to the ways you move. You’ll start to get a physical feel for the limitations of the devices — and with it, you’ll start to get ideas for how to expand their capabilities. The most common way I decide what to work on next is that I find myself thinking “I wish I could detect that” after I make a particular gesture a few times while rehearsing.

Build an instrument, not a toy

Most people who are building wearables are out on the fringes of experimental music and thus not especially interested in performing the same song over and over. However, I feel pretty strongly that even if this isn’t your goal, you should still try to design a device where you at least have the option to write a song and then perform it more or less the same live. In my opinion, this is what differentiates a wearable instrument from a toy. An instrument is something that you can use to reliably reproduce the same sounds with the same set of gestures, meaning you can learn to use it better over time.

Device Sensitivity

Another important part of using an instrument is choosing the level of control you have over it. In a nutshell, anything you build needs to have some sensitivity to external inputs (if it doesn’t, why are you building this as wearable tech and not just a box you can put on a table?), but it can’t be either so sensitive that you lose control over it, or so insensitive that you can’t tell that you’re doing anything.

The Z axis on the Scalar Glove has a 90-degree range that corresponds to 0-127 in MIDI control change — when my hand is parallel to the ground, the CC value is at 0, and when I rotate it ninety degrees so it’s perpendicular to the ground, it’s at 127. This is nice because it gives me control, but it’s also pretty intuitive — I rotate my hand and it directly corresponds to a detectable sweep in tone (or distortion level, or whatever). The X axis isn’t as good for this because it has a 360-degree range of motion and I have to move my whole arm — essentially turn in a circle — to access that whole range, which isn’t practical; it’s on my to-do list to make that more useable.

Appearance and Structure

mask_inside
The glamorous inside of my latest mask, before I did all the wire control. You can see the equally-glamorous drawer non-slip liner I’m using for padding.

If you want to use your wearables in a performance context, how they look does matter. Tech nerds like me will be excited no matter how they look as long as they work in an interesting way — but for the majority of audiences, the novelty will wear off really fast if they look bad or you look awkward and ungainly while using them. I recommend starting with a general aesthetic in mind, and on that front, my advice is this: wearables are inherently “futuristic”, so you don’t need to go all in on the sci-fi aesthetic. If you’re wearing some weird crap covered in wires, you will look like a cyborg; it’s inescapable. So in general I’d recommend staying away from the chrome, lazers, optical filaments, etc.

There are a lot of ways to build the structure of your wearables — that is, the non-electronic, physical part of the devices. If you have access to a 3D printer or a CNC machine, obviously those are some of the better tools for the job. If you don’t — and I never have — there are materials and tools that you can work with in your own home that can still look really great. My earlier versions of the gloves used a lot of cross-stitching plastic canvas. It’s extremely cheap at any craft store, and it’s soft enough that you can easily cut it with scissors but rigid enough to use as a mount for your electronics. It’s also breathable and it has all these little holes that are perfect for wire management.

Lately I’ve been getting into thermoplastics. They’re really fun to work with and I really recommend it. If you don’t have a CNC machine or any other subtractive manufacturing tool for cutting plastic, I can’t recommend strongly enough that you get a Dremel. You can spend less than $100 to get a full kit with a set of bits, and my Dremel changed my life. You can use it to cut out really intricate patterns, by hand, with very good control. It’s fantastic.

If you use thermoplastic, you’ll probably want to sand it before you paint it, because otherwise the paint may not stick. Some types of spray paint bond really well with plastic, but even those can sometimes get scraped off through regular wear and tear.

You’ll probably want some kind of padding between the devices and your skin. Personally, I use that drawer liner stuff. It’s soft and breathable and adds a nice amount of padding without making the device bulky, and it holds well with a little bit of glue (or you can even tie or sew it in place since it also has little holes). It’s also super cheap — and you can even use the rest of the roll in your kitchen drawers.

These days, any time I’m walking through a supermarket, hardware store, craft store, or honestly almost anywhere, I have this little part of my brain that evaluates the stuff I see for its weird instrument capabilities. Almost everything out there has some potential use that’s more interesting than what it’s intended to do.

Thanks for coming to my TED talk

I could write another thousand words about this, but I’m gonna stop here. If you found this useful, I’d love to hear from you through the contact form, or on Facebook, Instagram, or Twitter.

Thanks,

80KV

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Powered by WordPress.com.

Up ↑

%d bloggers like this: