Interferator M4L — Behind the Device

There’s a specific frustration I kept running into when building patches in Ableton: modulation sources that were either too random or too mechanical. A plain noise LFO feels arbitrary — it moves, but it doesn’t mean anything. A regular LFO is predictable to the point of becoming invisible. What I was looking for was something in between. Something that felt alive without being chaotic, structured without being obvious.

That’s where Interferator started.

The Problem with Random

I like semi-random modulation. Noise is useful as a modulation source, but when there are algorithms working underneath — Perlin, Voronoi, Simplex — it starts to feel more organic. The movement has coherence even when you can’t quite explain why. It follows its own logic.

But there’s a line. Push too far toward pure randomness, and you lose the musical thread entirely. Too far the other way, and you’re back to a machine gun pattern or a predictable LFO curve. What lives between those two extremes is interesting territory — and that’s what I wanted to build a tool for.

(There’s a whole other conversation to be had about the grey zone between machine gun and motorik in the Krautrock sense, but that’s a different device.)

Four Layers, Three Algorithms, One Field

A single noise field felt too simple. From the beginning I wanted Interferator to respond musically — not just spatially. So each of the four layers has its own rhythmic relationship to the tempo. You set the zoom level independently per layer, meaning each one moves at a different rate relative to the beat. They can lock in with the groove, drift against it, or anything in between.

The layers aren’t just stacked — they’re combined using three different algorithms that create interference patterns between them. This is where the name comes from. What you end up with isn’t a noise generator. It’s a field. A living two-dimensional space where the patterns that emerge are a result of those interferences — and they can follow the rhythm, or not. They just are.

Probes: Sampling the Field

One sampling point would have been enough to generate modulation. But a single point puts you back in familiar LFO territory — one signal, one destination.

Three Floating Probes change that completely. You place them freely in the 2D field, and each one outputs its own modulation signal — all drawn from the same underlying interference field. The key insight is spatial: probes that are close together will output signals that are related. Not identical, but correlated. They’re reading the same neighborhood of the field.

Map two nearby probes to filter cutoff and delay feedback. Suddenly those two parameters aren’t moving independently — they’re moving together, because they share a spatial context in the field. That relationship isn’t programmed. It emerges from where you put the probes. It’s one of those behaviors that’s hard to describe but immediately obvious when you hear it.

Taming 56 Parameters

Under the hood, each layer has more controls than what’s exposed in the UI. Early builds were unusable — not because the concept was wrong, but because there was simply too much to manage. Too many parameters, too many decisions, too much friction between an idea and hearing it.

The breakthrough came from an unexpected direction: I started designing Interferator for Push. Not for performance, but as a design constraint. Push gives you eight parameters per page. That limit forced the question: which eight parameters actually matter per layer?

Getting there took a while. I worked through the reduction iteratively — often using an LLM as a sparring partner to think through tradeoffs. Not to generate ideas, but to argue through them. When we’d worked out a direction, I’d implement it, try it, and come back with what didn’t feel right. Some of that back-and-forth nudged parameters around, renamed macros, reconsidered what belonged on the surface and what could stay underneath.

What’s left is eight controls per layer: enough to shape the character of each independently, not so much that you drown before you find something useful.

What It’s For

Honestly? I’m still finding out.

The obvious use is textural modulation — pads, sustained sounds, arpeggios. With arpeggios in particular, the arpeggiator itself does the sampling: each new note picks up whatever the probe is reading at that moment, so the modulation arrives in rhythmic slices without Interferator ever stopping. The field is always moving. It’s never holding.

But I’ve also mapped it to the macro controls of an effect chain and gotten results I didn’t expect. And to the parameters of a Drum Rack — suddenly the kit has a slow, correlated drift to it that makes the loop feel like it’s breathing.

That last part still surprises me sometimes. I think that’s a good sign.

---

Interferator M4L is available now at phasenpunkt.gumroad.com. Requires Ableton Live 12 or higher with Max4Live.