NOMN

A sound in space is never spectrally uniform: different frequencies arrive from different directions and reflect off different surfaces, and each decays at its own rate. The perception of spatial life in audio, the sense that a sound occupies a place, arises from these spectral asymmetries. Acoustic environments produce them automatically, and convolution and spatialization tools reproduce and model them; CHAKRABOOST lets you compose with them as raw material for effects that have no acoustic equivalent.

The multiband phaser is an old topology, which splits a signal into frequency bands, phase-shifts each one and recombines them. CHAKRABOOST treats each band as a complete processing channel with its own phaser, compressor and position in space, all driven by the same modulation source, so the spectrum moves to one shared rhythm while each band responds to it in its own way.

There are 8 frequency bands, each with a vactrol-modeled phaser, an adaptive parallel compressor and a spatial panner with modulated orbits. The low end can sit still while the upper mids sweep left and the air frequencies orbit overhead, all pulsing with one rhythm, or each band can be set to move on its own.

The default modulation source is a small neural network, a learned prosodic-trajectory model from NOMN's microtiming research repurposed as a modulation source, which generates continuous, non-repeating curves that move the way speech and song do. Phaser sweeps, pan orbits and compression timing all run on these trajectories. All 8 bands share the same underlying motion, with a per-band amount of noise added so the bands stay correlated while each drifts slightly from the others. Any band can instead run a tempo-synced or free-running LFO.

At low dry/wet settings (15 to 30%) with Pan Depth at zero, CHAKRABOOST adds a subtle multiband phaser character and leaves the stereo image where it was. Pushed further, it gives a static pad internal motion and a 3D position, makes a drum bus phase and breathe across the spectrum, and makes a vocal sound as if it were being transmitted through eight different rooms at once.

With Pan Depth at zero on every band, CHAKRABOOST preserves the stereo image end to end and folds cleanly to mono. Pan Depth defaults to 0.5, so motion is on when the plugin loads. Per-band pan modulation moves frequency content across the stereo field, which causes phase cancellation in mono, so output that has to survive a mono fold should either keep Pan Depth at zero or run CHAKRABOOST on a send.

Audio enters through a true stereo crossover that splits each channel into 8 phase-coherent frequency bands across 20 Hz to 20 kHz, with crossover points at 80, 160, 320, 560, 1120, 2240 and 4200 Hz by default (each of the seven is adjustable through host automation). The global Phase Mode parameter switches between two crossover topologies.

Zero Latency (default) uses IIR Linkwitz-Riley 4th-order filters in a balanced binary tree with allpass phase compensation, which sum flat in magnitude.

Linear Phase uses a 255-tap windowed-sinc FIR with a Blackman window, which sums flat in both magnitude and phase. In this mode the vactrol phaser is the only source of phase manipulation in the signal, and every phase relationship in the output follows from the phaser's settings.

Each band passes through three processing stages in series.

Phase modulation. Each band runs an all-pass phaser of 2, 4, 6 or 8 stages (4 by default), with every stage controlled by an independent vactrol model, and the phaser's notches are anchored to just-intonation ratios within the band. The vactrol responds asymmetrically and carries thermal memory, and the per-band Memory control sets how slowly it moves, from a 0.5 ms attack with 30 to 200 ms release at one end to a 20 ms attack with 600 to 3000 ms release at the other. Density, Resonance and an LFO Invert switch complete the phaser controls, and tanh soft-clipping in the feedback path allows resonance without runaway.

Adaptive compression. Each band has parallel (NY-style) compression with timing driven by crest factor. The Adaptivity control (0 to 1 per band) sets how far the timing scales with program material: at 0, attack and release are fixed, and at 1 they vary by a factor of 4 with the signal's dynamic character. This keeps each band's level stable as the phaser reshapes it, and the compression stays out of the modulation's way.

Spatial output. The global Spatial Mode parameter selects how each band reaches the output.

- Stereo: each band is panned through a two-speaker VBAP layout with mid/side decoding, so the mid moves while the side is preserved. Dragging the two speaker positions in the layout editor warps the stereo image.
- Surround: VBAP positioning across any output layout from 2 to 32 channels, with presets for Quad, 5.1, 7.1, 5.0.4, 7.1.4 and 9.1.6, a catalogue of ring, cube, box and dome geometries, and custom layouts. Each band orbits in X and Y, and an Orbit Angle rotates the orbit's path.
- Headphone Preview: each band becomes a point in 3D space with its own X/Y/Z position, rendered with interaural level and time differences (the far ear delayed by up to 0.7 ms) for headphones. Each band is summed to mono in this mode, since each one is a single point source.

The SPATIAL tab holds the speaker layout editor, an isometric plot where speakers can be dragged, typed in numerically, added, deleted, saved and recalled, with an activity ring on each output. Each output can be assigned a Full, Small or Sub role, and bass management sends everything below the Bass Xover frequency (40 to 200 Hz) to the sub outputs at the Sub Level. A listener card sets the listener's X/Y/Z position, which can also be dragged in the plot, and a global Spread control widens each band's image across neighbouring speakers.

The modulation for all of these comes from the NeuralLFO, a single learned prosodic-trajectory model that all 8 bands read from. Per-band Rhythm Strategy and Articulation controls steer where each band sits in the model's latent space, per-band Temperature sets how far each band wanders from the shared trajectory, and Neural Depth sets how strongly the trajectory drives the band. A background worker computes the trajectory ahead of playback, and offline renders compute it in line. The Bar Sync button restarts the motion on the host's bar boundaries. Through host automation, each band's Mod Source can be switched from NOMN to Tempo Sync (12 rates from 1/1 to 1/16 triplet) or Free LFO.

Any modulatable knob can also be driven from the sidechain input. Right-clicking a knob assigns, edits or removes a sidechain routing with a depth from -1 to +1 and optional polarity inversion, and an SC readout in the MIX tab shows the sidechain envelope. Routings are saved with the session and with presets. The same right-click menu can lock a band's pan modulation in place.

All bands recombine, and a global dry/wet mix blends the processed result with the original signal. A Master Intensity control scales phaser depth and mix, pan depths, neural depth and compressor mix together across all bands. Input Gain and Output Gain each trim by up to 24 dB. Output protection is set by Safety Mode, which defaults to a lookahead limiter (1.5 ms lookahead, ceiling at -0.92 dBFS) with tanh soft clipping behind it, and can be switched to soft clipping alone or turned off through host automation.

On a synth pad or drone. Slow modulation, moderate Neural Depth and wide spatial orbits give the pad internal motion, with different harmonics drifting to different positions, and a three-dimensional quality that stays below the threshold of an obvious effect.

On drums or percussion. A faster modulation rate, low feedback and subtle spatial depth add a rotating, breathing quality to the transient tail. With low Adaptivity for predictable timing, the compressor keeps the attack intact while the phaser works on the sustain. A sidechain routing from the kick or snare can push the motion in time with the groove.

On a recorded vocal. Gentle phaser depth on the mid and upper bands only, with light spatial movement, gives a diffuse, spatialized quality without the obvious swoosh of a traditional phaser. Linear Phase mode helps here, since it leaves the vocal's phase character shaped entirely by the phaser.

On a full mix bus, with caution. Linear Phase mode, dry/wet at 10 to 20%, Pan Depth at zero, slow modulation, low Adaptivity and a low per-band phaser mix add micro-variation across the spectrum without smearing the image or breaking mono. Output Gain compensates for the level of the per-band processing.

For headphone-first delivery. Headphone Preview places each band at its own X/Y/Z position, and the bands move through 3D space along the same shared trajectory. The effect is strongest on headphones and narrows on speakers, because each band is summed to mono in this mode.

For immersive mixes. On a surround or height bus, each band can orbit its own region of the room, and the bass management and per-output roles let the same session play on a full-range rig or one with small speakers and subs.

As a sound design tool. Fast modulation, high feedback, 8-stage phasers, wide orbits and aggressive compression can turn a field recording into something unrecognizable, with separate phaser, compressor and 3D position settings for every frequency range.

In Stereo mode the processing is true stereo, with independent crossover banks and phasers for the left and right channels, driven by shared modulation, and stereo-linked compression detection (gain reduction computed from max(|L|, |R|) per band and applied identically to both channels). There are 8 parallel processing chains with per-band solo and mute, and every parameter on every band is automatable in the DAW.

CHAKRABOOST reports a fixed latency to the host in every Phase Mode, covering the linear-phase crossover (127 samples), the limiter's lookahead and any speaker distance compensation, and the DAW compensates for it automatically.

The visualization is an 8-lane scrolling waveform display, one lane per band, showing about 5 seconds of the real post-processing amplitude envelope, time-aligned, with pan position drawn as vertical displacement within the lane. Clicking a lane selects that band for editing, and each lane carries solo and mute keys, with a key to clear all solos. Each band also shows a polar plot of its vactrol's behaviour, a scope of its modulation waveform, a stereo level meter and a gain-reduction meter. Multichannel input and output meters, a limiter gain-reduction display and a NeuralLFO status indicator sit in the master strip. The interface is greyscale, with bands told apart by brightness.

Band settings can be copied from any band and pasted to another band, to all bands or to a chosen band, and the last paste can be undone.

Presets are saved as named .chakrapreset files in ~/Library/Chakraboost/Presets/ and recalled from the dropdown at the top of the window. Saving under an existing name asks whether to update it or save a new one, so nothing is overwritten silently. CHAKRABOOST ships with no factory presets. Signing in with a nomn.jp account syncs presets across devices and opens a community library for sharing and browsing other users' presets.

Global: Input Gain, Output Gain, Dry/Wet, Master Intensity, Bypass, Phase Mode, Spatial Mode, Spread, Listener X / Y / Z, Bass Xover, Sub Level, Bar Sync, Soft Clip, Safety Mode, Crossover 1 to 7.

Per-Band Phaser (×8): Depth, Feedback, Phaser Mix, Rate, Phaser Stages, Phaser Memory, Phaser Density, Phaser Resonance, LFO Invert.

Per-Band Spatial (×8): Pan, Pan Y, Pan Z, Pan Depth, Pan Depth Y, Orbit Angle.

Per-Band Compression (×8): Comp Threshold, Comp Ratio, Comp Attack, Comp Release, Comp Mix, Comp Adaptivity.

Per-Band Modulation (×8): Mod Source, Sync Rate, Rhythm Strategy, Articulation, Neural Depth, Temperature.

Per-Band Utility (×8): Gain (±18 dB), Solo, Mute.

Safety Mode, Crossover 1 to 7, Mod Source and Sync Rate are available through host automation only.

Key Features:
- 8-Band Spatial Processing: full-spectrum frequency splitting with a phaser, compressor and spatial panner on every band.
- Two Crossover Topologies: Zero Latency IIR LR4 (default) or linear-phase FIR with a flat phase response.
- Vactrol Phaser Modeling: LED/LDR optocoupler simulation with asymmetric attack and release, thermal memory, 2 to 8 stages and a per-band Memory control.
- NeuralLFO Modulation: a shared learned prosodic-trajectory model from NOMN drives all 8 bands with continuous, non-repeating curves that carry the timing character of speech, with tempo-synced and free LFOs available per band.
- Sidechain Modulation: any modulatable knob can follow the sidechain input, with adjustable depth and polarity.
- Adaptive Parallel Compression: crest-factor-driven timing with a per-band Adaptivity control and an NY-style parallel mix.
- Three Spatial Modes: Stereo (VBAP with mid/side decoding), Surround (VBAP on any layout from 2 to 32 channels, including 9.1.6), and Headphone Preview (per-band X/Y/Z position with interaural level and time differences).
- Speaker Layout Editor: drag-and-drop speaker and listener placement, layout presets, Full / Small / Sub output roles and bass management.
- Output Protection: a lookahead limiter with soft clipping behind it, on by default.
- Presets: user presets with overwrite protection, device sync and community sharing through a nomn.jp account.
- Full DAW Automation: every parameter across all 8 bands.

Compatibility:
Formats: Mac VST3 and Audio Unit (AU); Windows VST3 and Standalone.
OS: macOS 10.15+ (Apple Silicon and Intel), Windows 10 or later (64-bit).

Includes a 14-day free trial. After the trial, audio mutes until activation.