DSP

The synthesizer is just a set of delays, albeit very short ones, all feeding back into eachother, and at lengths that are not divisible by eachother (i.e. enharmonic). So one gong might have 6 delays, of length 500, 224, 102, 44, 38, 29. Because they don't factor into eachother nicely, they make dissonances associated with gongs.

It's a little bit like karplus strong or waveguide synthesis if that means anything to you.

When you touch, the delay lines are injected with a thump sound that I recorded by hitting my desk, recorded into my iPhone. This thump sound, echoes through the delay lines until it finally decays away, which can take a while(!)

To get different tones, I just play back that initial thump sound at a different speed (and pitch), which ends up exciting different delay lines more or less than others, sounding reasonably organic.

This is all fed into a reverb effect to space it out a bit more, then a compressor to manage any massive accidental peaks.

The audio code was programmed in C++, then converted to WASM using emscripten. The UI was all in javascript though. I made this repo to show the workflow of making audio dsp in C++ and convert it to a web page - it's mostly just a reminder for me to show how to set it all up again if I want to do another project like this.

Visuals

Not much to say here, it's WebGL, just a fullscreen quad getting drawn to by a glsl frag shader.