An intriguing mouth-played instrument emerged—and won—at the 2023 Guthman Musical Instrument Contest hosted by Georgia Tech. [Keith Baxter] took notice and reproduced the idea for others to explore. The result is the Zen Flute Mouth Theremin, a hybrid of acoustics, electronics, and expressive performance.
At its core lies a forced Helmholtz resonator, a feedback system built with a simple microphone and speaker setup. The resonator itself? The user’s mouth. The resulting pitch, shaped by subtle jaw and tongue movements, is detected and used to drive a MIDI controller feeding an external synthesizer.
Like a trombone or classic electromagnetic theremin, the Zen Flute doesn’t rely on discrete notes. Instead, the pitch is bent manually to the desired frequency. That’s great for expression, but traditional MIDI quantisation can map those “in-between” notes to unexpected semitones. The solution? MIDI Polyphonic Expression (MPE). This newer MIDI extension allows smooth pitch transitions and nuanced control, giving the Zen Flute its expressive character without the hiccups.
Physically, it’s an elegant build. A flat speaker and microphone sit side-by-side at the mouth end, acoustically isolated with a custom silicone insert. This assembly connects to a length of clear PVC pipe, flared slightly to resemble a wind instrument. Inside, a custom PCB (schematic here) hosts a mic preamp, an audio power amp, and a Teensy 4.1. The Teensy handles everything: sampling the mic input, generating a 90-degree phase shift, and feeding it back to the speaker to maintain resonance. It also detects the resonant frequency and translates it to MPE over USB. A push-button triggers note onset, while a joystick adjusts timbre and selects modes. Different instrument profiles can be pre-programmed and toggled with a joystick click, each mapped to separate MIDI channels.
Mouth-controlled instruments are a fascinating corner of experimental interfaces. They remind us of this Hackaday Prize entry from 2018, this wind-MIDI hybrid controller, and, of course, a classic final project from the Cornell ECE4760 course, a four-voice theremin controlled by IR sensors.
The Japanese did it first: EWI (Electronic Wind Instrument)
https://www.youtube.com/watch?v=TCei7ADE87Q
Can fit about any instrument in there for which you input samples.
The EWI can output MIDI, so you are not necessarily constrained to using samples.
EWI’s are pretty cool but the difference here isn’t in the sound it’s in the control. EWI’s are basically keyboard instruments. The Zen flute has only one button.
Reminded me of taking steel guitar:
https://youtu.be/_R9an8AU3No?feature=shared
Is the driver speaker loud enough to feel in your mouth? Seems like that would allow for a natural feedback in the musician, as opposed to them closing the loop on the sound that comes out of the device some many ms later and processed. This would give real physical feedback to the user on actual native pitch of the signal regardless of further processing and intrinsic delays.
I don’t feel vibrations in my mouth when I whistle; feedback is via the ears there too. Just because a computer is involved doesn’t mean there’s any appreciable human-scale latency; MIDI instruments would have all been dead in the water otherwise.
You do feel feedback if you use a reed instrument though.
As for whistling. you feel your mouth’s movement and muscle modulation, and even if you aren’t directly consciously aware.
It’s quite amazing how many kinds of proprioceptors we have and how much we rely on them without even realizing.
i think you just don’t know you can feel a whistle from the inside of your mouth
i think this is an interesting project because it straddles the boundary between real time innate feedback to the user (the resonance inside your mouth) and quantized / delayed midi output. midi guitars straddle the same boundary but i’m more skeptical of them because the guitar is an instrument in its own right. so the player will definitely ‘feel’ different about it, and it seems to me the difference will be largely inferior once the novelty wears off.
That’s not a “theramin” (properly spelled Theremin); it’s a Jew’s harp with extra steps.
I disagree, a jew’s harp is held between the teeth and then ‘twanged’
This just detects the shape of the inside of your mouth and resonates to that.
So it’s closer to a theremin than a jew’s harp IMO, but neither comparison really applies that well.
It is however quite an original approach to making an instrument.
Sort of a ‘breathless flute’ I’d say.
the jews harp definitely just detects the inside of your mouth and then resonates to that :)
psst: *Ther/e/min.