Very few genuinely new synthesis methods have emerged in the past twenty years. FM, granular, wavetable, physical modeling of hitting and bowing things, the paradigms were established in the early 90s. Since then, most "new" plugins are refinements, combinations, better interfaces, or indexes of existing ideas.
FOAM is an attempt at opening new territory.
Bubble acoustics is one of the most sonically rich phenomena in nature and one of the least explored in synthesis. The physics are well-documented but undertheorized as a sound source, largely because the behavior is so chaotic that nobody has bothered to model it in real-time.
The sound of a wave receding is not five or ten sources. It's thousands of interactions across a wide range of behaviors whose aggregate creates the sonic richness of the texture. FOAM implements this physics directly. Not samples. Not approximations. Actual acoustic modeling of what happens when gas meets liquid, from the initial film rupture through cavity resonance to viscous damping.
It is a complex tool, and we recommend using it with automation curves for maximum detail and morphology. It is also a really fun performance synth you can play with MPE or drive with a modular setup. For noise enthusiasts there is an easter egg you'll need to find.
--
## How It Works
Each bubble is a physics equation, not a recording or grains.
FOAM models 11 distinct acoustic stages of a bubble's life, from the initial membrane rupture through resonance, collapse, and decay. Each stage is governed by its own physical behavior, and every parameter maps to real fluid dynamics.
Every parameter feeds the acoustic formulas. Turn up viscosity and the math changes: small bubbles die fast and large ones ring longer in thick liquids, as they do in nature, as best is known within current bubble acoustics research.
At peak density, thousands of synth voices render simultaneously across multiple CPU cores,
producing textures with a level of temporal detail that simply isn't
possible to similarly shape for creative use with sample playback or wavetable synthesis.
Because of the way the bubbles are calculated, for the most immediate,
transient-rich response, we recommend a buffer size of 64-128 samples. When CPUs are much much faster it will be possible to normalize the behavior regardless of buffer, but for now this is the best compromise between the scale of texture we can create and making it impossible to actually use FOAM within an OS.
--
## The Engine
Up to 16,384 simultaneous voices with cascade modeling, where parent bubbles trigger chains of interacting child bubbles. A configurable CPU governor lets you set the ceiling for voice density — open it up fully on modern hardware, or dial it back in dense sessions. Population dynamics create collective frequency shifts as density increases, and the transition from discrete pops to continuous foam texture emerges naturally from the sheer number of independent physical interactions.
The range runs from single droplets to dense foam rafts. From water to honey to mercury. From delicate rainy sounds to walls of noise. Almost realistic rivers if you want them. Total textural exploration if you don't, with fantasy-sized bubbles that might also give you a kick or sub-bass if you can find it.
--
## Parameters
Bubble Physics
- Rate: Emission rate and energy.
- Size: Base bubble size (pitch).
- Ring: Ring time / decay character.
- Cascade: Child bubble chain depth.
- Glide: Portamento between notes.
Realism
- Shimmer: Surface ripple modulation.
- Natural Balance: Film rupture vs cavity collapse mix.
- Jets: Secondary ejection modeling.
- Mirror Depth: Submersion depth and surface reflection.
Texture
- Stereo Width: Stereo spread.
- Magic: Psychoacoustic intensity (phantom sub-bass, cochlear distortion, binaural roughness).
Parameter Modulation
- Chaos: Parameter randomization with per-knob locking.
- Drift: Slow Brownian random walk across parameters.
- Scatter: Per-voice randomization.
Liquid Identity
10 Liquid Material Presets: Water, Seawater, Soap Solution, Glycerol, Honey, Olive Oil, Mercury, Ethanol, Molasses, Custom.
- Liquid Mass: Liquid density (manual).
- Tension: Film strength (manual).
- Viscosity: Damping character (manual).
Foam Topology
- Topology: Structural foam simulation (coarsening, collapse events).
- Foam Age: Temporal evolution of foam structure.
- Wetness: Liquid fraction in foam.
- Stress: Mechanical stress on foam network.
- Yield: Yield stress threshold.
- Jamming: 2D to 3D structural transition.
- Burstiness: Event timing statistics (regular vs clustered).
- Aging Rate: Speed of structural evolution.
Surge Modulator
- Surge: Temporal envelope shape (Wave, Pummel, Roil, Gush, Burst).
- Surge Time: Envelope duration (1–30 seconds).
- Surge Depth: Modulation intensity.
Froth Gen
6 emitter modes controlling how bubbles are released over time:
- Poisson: Natural random emission.
- Nucleation: Carbonation-style release.
- Foam Raft: Avalanche dynamics.
- Turbulent: Intermittent turbulence.
- Melodic: Pitch-stable emission for tonal playing.
- Harmonic: Scale-quantized emission for harmonic content.
Pitch Gravity
Bubbles generate their own resonant frequencies based on size, but for creative application you can pull these toward pitch sets.
- 76 Scales: From traditions around the world, plus microtonal systems, xenharmonic tunings, and synthetic constructions.
- Root: Transpose to any key.
- Quantize: Scale conformance strength.
Voicing Dynamics
When thousands of bubbles sound simultaneously, how they sum changes the character of the sound. FOAM's mix section gives you control in two stages.
- Gain Scaling: How bubble amplitudes are balanced as population grows.
- Freq Summing: Whether low and high bubbles contribute equally or scale by physics.
- Spatial: Spatial position influence on voice amplitude.
- Transients: How much the initial pop attack is preserved through scaling.
- Drive: Pre-compressor input gain.
- Multiband Comp: 8-band parallel compressor (63Hz–8kHz) with per-band mix and adaptive dynamics.
--
## Features & Formats
Key Features:
- Full MPE Support: Per-note pitch bend, slide, and pressure for expressive control over bubble size, pitch sweep, ring time, emission rate, and amplitude.
- GPU-Accelerated Visualization: Real-time display of bubble populations, density, and spectral content.
- MIDI Learn: Right-click any parameter to assign CC control.
- OSC Output: Stream bubble events to external applications for synchronized visuals or data sonification.
- Preset System: Save and recall configurations. Preseeded with statistically distinct timbres, not bespoke ideas. It's an open terrain for your own exploration.
Compatibility:
Formats: VST3, Audio Unit (AU), Standalone.
OS: macOS (Apple Silicon + Intel) and Windows (experimental as of April 2026).
--
## A Note on Block Size
FOAM is a stochastic physical model. Every bubble is spawned by processes
governed by fluid dynamics equations, and thousands of these interact per
second. Your DAW's audio buffer size will affect the output. Like an acoustic
instrument, FOAM's character responds to the conditions it plays in.
This is not a bug. It is the nature of modeling chaotic physical systems in
real time at their actual scale vs grains or wavetable.
At lower buffer sizes, the engine resolves finer temporal detail — more
transient definition, more continuous texture.
For the most immediate response, set your buffer to 64 or 128 samples. This
may complicate sessions built around large sample libraries that benefit from
higher buffer sizes. If you need to work at 512 or 1024, the sound will be
different — not necessarily worse for your case, but different.