BassDroplet

A breath-played instrument that drips water onto metal. Same contest as WindGate, a year later, with no computer in the sound path.

Tangible Music Lab - Out of Control 2026

A celebration of musical instruments, their technologies, and the bodies that play them, organized by the Tangible Music Lab, University of Arts Linz

Kunstuniversität Linz

Video by Ulla Rauter.

The brief

The Tangible Music Lab at Kunstuni Linz runs Out-of-Control, a contest about instruments, the technology inside them, and the bodies that play them. The rule holds from year to year: build something you can play with anything except your hands. I went in 2025 with WindGate, a mask with nine infrared sensors behind a grid of flaps that you play by breathing at it. When the 2026 call came round I was glad to have a reason to build something new for it.

Breath again, aimed somewhere completely different.

WindGate is a controller. Nine clean channels of breath data leave the mask and land in a patch on my laptop, and the thing making the sound is that patch, not the thing on my face. I'd been curious for a while about the other end of that: breath in one side, something physically happening on the other, and nothing in between that has to boot.

Water falling into a metal bowl.

How it works

You blow into a tube. A suck-and-puff sensor, the kind used in accessibility hardware, reads pressure in both directions, so blowing and drawing back are two separate gestures instead of one axis with a dead middle.

That goes to an ESP32, and the ESP32 drives a 3D printer control board left over from an earlier build. Printer boards are a good deal for this. Stepper drivers, screw terminals, a power stage, all on one board and all sized for the motors I wanted to turn. Half joke, half the obvious answer.

The steppers turn peristaltic pumps. The pumps push liquid up a printed gooseneck arm, and it drips into a steel bowl.

The pump

Peristaltic, because nothing else does all three of the things I needed: small amounts on demand, an instant stop, and no pump internals in contact with the liquid. A rotor with bearings as rollers squeezes a silicone tube against the inside of a housing, so the tube is the only wet part. It's also simple enough to print.

First housing test prints, transparent PETG.

Then a second, tighter version.

Then an evening moving water back and forth between two drinking glasses, which looks ridiculous and is the only way to find out whether the flow is repeatable enough to play.

The bowl

Stainless steel. Metal because a drop landing on metal gives you a real attack and a long ring instead of a knock, and a bowl because the curve gathers the liquid where the drops land.

I drilled the centres on the drill press in the workshop to take a brass hose barb.

The bowl sits in a printed pedestal on a block of packing foam. Without the foam the table joins in, and the contact mics can't tell the difference.

Everything on the control side. The black box holds two channels of gain for the contact mics.

Nothing between the drop and the speaker

Two contact mics on the bowl, into the preamp, into a small Behringer mixer, a bit of EQ, out to the PA. That's the whole chain. Nothing anywhere in it analyses anything.

The 2026 call asked for MIDI notes and CC values, and BassDroplet doesn't send any. All the digital parts sit on the control side, between the tube and the motor, and the sound never touches a computer. That's a liberal reading of the rules and I'd do it the same way again, because the drop is already a good sound: a hard transient, and a pitch that keeps sliding down as the bowl fills. Anything I put in front of it would have been my interpretation of a sound that was managing fine without me.

Playing it turned out closer to a wind instrument than I expected. Steady pressure gives you a slow pulse. Push harder and the drops come faster and land harder, the level rises, the pitch falls. Stop, and the last few are still in the air on their way down.

Where it is now

The two instruments have opposite strengths, which I didn't plan. WindGate reads breath in fine detail and then hands the sound over to a patch. BassDroplet makes its own sound and has almost no control resolution: one pump, one bowl, breath moving the rate and not much else.

Two bowls of different sizes would give it intervals instead of one voice, so that's the obvious next thing to try. The suck side of the sensor is also doing nothing at the moment (besides stopping) and I don't have a good idea for it yet. Whether any of that turns it into a real instrument, or just into a second thing to mop up afterwards, I don't know.

Specifications

Playing it

June 2026, at the Tangible Music Lab. The jam ran outdoors on the metal grating with everybody joining in with their post-digital controllers and instruments.

Running on site. The water was colored with black food coloring to make it visually more appealing.

An evening at the TFL yard full of people playing things with everything besides hands.

Built for the Tangible Music Lab's Out-of-Control Contest, part of the Postdigital Lutherie programme at Kunstuni Linz, 2026.