What is a soliton in audio?

Physical model

Dispersion spreads a wave. Nonlinearity pushes back.

A localized wave packet contains many frequency components. In a dispersive medium, those components travel at different speeds. The packet broadens. A sharp crest becomes a wider, softer shape because the parts of the wave no longer arrive together.

Nonlinearity adds the opposite tendency. The wave's own amplitude changes how it moves. In shallow water, a taller crest can travel faster than a lower one. In optics, an intense pulse can change the refractive index of the medium carrying it. Left alone, that amplitude dependence can steepen a wave until it breaks or produces unstable high-frequency structure.

A soliton sits in the useful middle. The dispersive tendency that spreads the packet and the nonlinear tendency that sharpens it stay matched. The wave persists because neither effect wins outright. That is the structural idea behind Soliton: build a feedback loop where every pass contains a spreading term, a shaping term, and a bounded energy target.

This boundary matters. Chiral's Soliton is not a claim that an audio feedback loop converges to the textbook sech-squared shape from the Korteweg-de Vries equation. The plugin is a discrete-time musical instrument with saturation, damping, a delay line, and safety limits. It borrows the mechanism that matters for sound: persistence through balance.

That phrasing is intentionally conservative. A physical soliton is usually discussed in a medium whose equations support a particular family of traveling-wave solutions. An audio delay is a designed feedback system. It has sampling rate, finite buffers, parameter smoothing, saturation curves, denormal protection, host automation, and user input. Those implementation details are not obstacles to the concept; they are the instrument. The credible claim is that the product translates the dispersion-nonlinearity-energy relationship into a stable audio control surface.

The equations

KdV gives the historical shape. NLSE gives the signal-processing bridge.

The Korteweg-de Vries equation was derived for long shallow-water waves in a rectangular canal. It is the equation that gave John Scott Russell's nineteenth-century observation a mathematical home. Its classic solitary-wave solution is a rounded pulse whose amplitude, width, and speed are tied together rather than independently chosen.

The nonlinear Schrodinger equation is the more direct signal-processing reference because it governs optical pulse envelopes in dispersive nonlinear fibers. Hasegawa and Tappert predicted stationary optical pulses in 1973; Mollenauer, Stolen, and Gordon observed optical-fiber solitons in 1980. The medium changes, but the architectural lesson holds: persistence comes from the relationship between dispersion, nonlinearity, and energy.

KdV and NLSE also explain why the word "soliton" is useful without being loose. The term does not merely mean "a wave that lasts a long time." It points to a specific kind of persistence: a localized structure survives because the forces that would normally destroy it are coupled. That is the conceptual filter for the audio design. A long feedback tail alone is not enough. The repeat needs a mechanism that changes its shape while keeping the loop bounded.

Audio mechanism

A normal delay has two boring futures: decay or howl.

A standard digital delay recirculates signal through a buffer with feedback gain. If the gain is below unity, each repeat is quieter than the last. If gain reaches or exceeds unity, the loop can accumulate energy until it clips, howls, or needs a limiter to rescue it.

Soliton adds structure inside the loop. The repeat is not merely copied and attenuated. It is transformed every time it returns.

The allpass chain creates phase-dependent delay without changing the static magnitude response. In time, that behaves like controlled dispersion: transients smear and the repeat broadens. Tone damping trims high-frequency energy before the nonlinear stage. Saturation, selected as Warm, Tape, Glass, or Fold, compresses and colors the waveform according to amplitude. The stabilizer measures the loop's energy and nudges it toward the Sustain target. The limiter is a backstop, not the musical engine.

The important point is not that the plugin solves a partial differential equation in real time. It does not need to. The point is that the controls map to real mechanisms: Spread moves dispersion, Focus moves nonlinear shaping, Sustain sets the target energy, Time sets the recirculation interval, Mix sets wet/dry balance, and Output sets final level.

This creates three practical regimes. Spread-dominant settings give the loop permission to diffuse: repeats become wider, softer, and more spatially ambiguous. Focus-dominant settings make the nonlinear stage more audible: transients tighten, harmonics step forward, and the repeat can feel more carved than smeared. Balanced settings are the center of the design. The repeat keeps being transformed, but the transformation has enough counterforce to remain musically stable.

The stabilizer is what makes that plane playable. Without it, a feedback delay near unity gain asks the user to balance on a knife edge. Too little gain and the line dies. Too much and the loop takes over the mix. With a target energy inside the loop, Sustain becomes a musical parameter rather than a dangerous feedback trim. It does not remove dynamics; it defines the reference point around which the loop breathes.

Interactive lab

Move the balance. Watch the pulse change state.

This lab is a readable sketch of the balance, not the Soliton source code. Spread broadens the pulse. Focus narrows and sharpens it. Sustain draws the target energy line. The most stable zone is where the spread and focus terms are close enough that the pulse breathes without flattening or pinching.

Control language

The front panel names the mechanism.

The four saturation shapes, Warm, Tape, Glass, and Fold, are not separate claims about soliton physics. They are musical transfer curves placed inside the loop. Each shape changes how the nonlinear term sounds as repeats stack and stabilize.

That distinction keeps the naming honest. Spread and Focus are mechanism controls. Warm, Tape, Glass, and Fold are timbral choices for the nonlinear stage. Sustain, Time, Mix, and Output are operating controls. The product becomes legible because each label tells the user what part of the loop is being moved.

Listening guide

Start with transients, then hold the line.

The clearest way to hear the mechanism is to send a short pluck or drum hit into the delay, then move through the Spread and Focus plane. Low Spread and low Focus behave close to a clean delay. High Spread with low Focus becomes broad and washed. Low Spread with high Focus becomes tighter and more saturated. Higher values of both can hold the repeat while still changing its internal shape.

Use Sustain as the energy target. Low Sustain lets the line fall away. Higher Sustain holds the repeat longer. Tone determines how much high-frequency material survives each pass, which matters because saturation will expose any excess brightness.

After transients, try a sustained pad or vocal phrase. Sustained sources reveal whether the loop is merely loud or actually organized. Raise Spread until the repeat stops feeling like a copy. Raise Focus until the center of the repeat returns. Then move Sustain last. If the line holds without flattening into a static drone, the balance is doing audible work.

References

Source material.

Related

Continue the wave trail.