Introduction

Particules is a granular texture synthesizer for VCV Rack. It takes a continuous stream of incoming audio, stores it in a short internal buffer, and plays back tiny slices of that audio — called grains — layered on top of each other, at different speeds, pitches, and positions in time. The result can range from subtle shimmer and blur, to lush sustained pads, to rhythmic stutters, to complete transformation of the source material into something unrecognizable.

It can also work as a delay, a beat slicer, and a standalone granular wavetable synthesizer — but granular texture processing is its heart.

Particules is based on the design of Beads, the classic granular module from Émilie Gillet of Mutable Instruments. The core of the code for Particules was written by Neal Sanche for No Such Texture, his adaptation of Beads for the Disting NT. Particules could not exist without their work, although they are not responsible for any bugs, errors, or poor design decisions in it.

What is granular synthesis?

Imagine you have a short recording of any sound — a chord, a drum hit, a voice, a field recording. Now imagine cutting that recording into hundreds of tiny fragments, each a few milliseconds to a few seconds long. You play back many of those fragments simultaneously, overlapping and layering them. You control where in the recording each fragment comes from, how long it plays, how fast (and in what direction), and what shape its volume envelope takes. The result is not a replay of the original — it’s a continuously evolving texture built from the material.

That’s granular synthesis. The fragments are grains. Particules creates up to 30 of them at a time.

Quick Start

Get a texture in one minute

1. Patch any audio source into IN L — a VCO, a sample player, anything with sustained content works well to start.
2. Patch OUT L to your mixer or output module.
3. Set DRY/WET fully clockwise (all wet).
4. Set DENSITY to about 11 o’clock (just left of center — constant grain rate, not random).
5. Set SIZE to about 2 o’clock (medium grain length).
6. Set SHAPE to 12 o’clock (balanced envelope).
7. Set TIME to 12 o’clock (midpoint in the buffer).
8. Send audio. You should hear your source material turned into grains.

From here, try these adjustments one at a time:

• Turn DENSITY further CCW — grain rate increases, staying steady and regular.
• Turn SIZE CW — grains get longer, texture gets denser and more sustained.
• Turn TIME slowly CW — you’re reaching further back into the recording buffer, creating a delay-like distance from the live signal.
• Turn PITCH slightly off 12 o’clock — new grains will play at the shifted pitch. Existing grains keep their original pitch until they expire, so the change happens gradually as the grain cloud turns over.

Adding some modulation

Once you’re comfortable with the basics, try this:

1. Set up the basic patch above.
2. Patch a slow triangle LFO into the TIME CV input.
3. Turn the TIME attenurandomizer (the small knob below TIME) to about 2 o’clock. The LFO now scans back and forth through the buffer, creating a slowly sweeping texture that moves in time with the LFO.
4. Add a second LFO — faster — into the PITCH CV input.
5. Turn the PITCH attenurandomizer to about 1 o’clock. Grains now shimmer with subtle pitch variation.

How It Works

Before going through the controls one by one, it helps to have a mental model of what Particules is actually doing. This section is that model — once you have it, the rest of the manual is details.

The recording buffer

Particules continuously records incoming audio into an internal buffer — a short loop of memory. The buffer holds between 4 and 32 seconds of audio, depending on the quality mode you’ve selected (more on that in Audio Input & Quality). It’s always recording, always updating, always replacing the oldest audio with new audio.

Think of it as a short tape loop, constantly recording. The record head is always moving forward; the playback heads (the grains) can read from anywhere on the loop.

When you engage FREEZE, recording stops. The buffer holds its current contents and grains continue to play from the frozen material.

What a grain is

A grain is a short playback event: a fragment of audio taken from a specific position in the buffer, played at a specific pitch and speed, shaped by an amplitude envelope, and then discarded. Particules plays up to 30 grains simultaneously.

The main grain parameters are:

• TIME — where in the buffer the grain reads from
• SIZE — how long the grain plays (and whether it plays forward or in reverse)
• PITCH — the playback speed / transposition
• SHAPE — the amplitude envelope

Creating a grain is called seeding. The DENSITY knob and SEED CV input control when and how often grains are seeded.

Parameters are snapshots

This is one of the most important things to understand about granular synthesis: when a grain is seeded, it takes a snapshot of all the current parameter values and holds them for its entire duration. If you turn a knob while a grain is playing, that grain is not affected — the change only applies to the next grains seeded.

This snapshot behavior is what gives granular synthesis its characteristic texture: add some randomization to PITCH, and each new grain picks a slightly different pitch and holds it. The result is a cloud of different pitches playing simultaneously, even though you only have one pitch knob.

How attenurandomizers work

Each grain parameter (TIME, PITCH, SIZE, SHAPE) has a smaller knob beneath it. These are attenurandomizers — a word that blends “attenuate” and “randomize.” They work differently from the attenuators or attenuverters that are common on modular devices, and they serve different purposes depending on whether a CV cable is patched into the corresponding CV input. The full details are in Attenurandomizers in the Controls Reference.

Controls Reference

Audio Input & Quality

Inputs and outputs

Particules has stereo inputs (IN L and IN R) and stereo outputs (OUT L and OUT R).

• Mono: Patch only IN L — Particules processes mono and outputs stereo.
• Stereo: Patch both inputs — Particules processes and outputs in stereo.
• If you leave OUT R unpatched, the left and right outputs are summed to OUT L.

Particules continuously records the incoming signal into its buffer as long as FREEZE is not engaged.

Input gain

Particules includes automatic gain control. When a cable is patched into an input, it monitors the incoming signal level for a few seconds and adjusts the input gain to a useful level (0 dB up to +32 dB, with headroom preserved).

If the automatic gain isn’t giving you the right level, you can override it from the context menu — see Context Menu.

Quality modes

The quality button cycles through four modes, indicated by the button’s LED color. These modes affect much more than just audio fidelity — they change the buffer length, the character of the feedback limiter, and the tone of the reverb.

• Bright Digital — pristine, clean. Brickwall limiting on feedback. Shortest buffer.
• Cold Digital — 12-bit character reminiscent of older granular processors. The dry signal still runs at 48 kHz even though the buffer is 32 kHz.
• Sunny Tape — warm, slightly soft character.
• Scorched Cassette — lo-fi wow and flutter. Low end swells, high end rolls off, adds noise and saturation. Longest buffer (up to 32 seconds mono). If you want tape delay vibes or maximum buffer length, this is your mode.

Quality mode changes made while FREEZE is engaged are deferred — they take effect when FREEZE is released.

Grain Generation

DENSITY

DENSITY controls how often grains are seeded — the rate of grain creation.

• 12 o’clock: No grains generated.
• Turned CCW from 12: Grains generated at a regular, steady rate. The further CCW, the faster the rate. At the extreme CCW position, grains are seeded at audio rate (the period of the note C3).
• Turned CW from 12: Grains generated at a randomly varying rate. The further CW, the faster the average rate, but timing varies unpredictably. At extreme CW, also reaches audio rates.

The left side of DENSITY gives you mechanical, regular grain clouds. The right side gives you something more alive and chaotic.

A white LED near the DENSITY input flashes on each grain trigger — useful for monitoring what’s happening.

SEED CV input

The SEED CV input accepts a trigger, gate, or clock signal to control grain creation externally. The behavior of the SEED CV input depends on the SEED CV mode setting in the context menu:

Triggers mode (default)

Patch a trigger or clock into SEED. Particules uses the rising edge of each incoming pulse to fire grains. DENSITY controls the relationship between the incoming pulses and grain firing:

• Turned CCW from 12: Clock division — a grain fires every Nth pulse (further CCW = less frequent, up to every 16th pulse).
• 12 o’clock: One grain per pulse (1:1).
• Turned CW from 12: Probability mode — each pulse may or may not fire a grain. Further CW = higher probability, up to 100% at full CW.

Gates mode

Switch to Gates mode from the context menu. SEED acts as a gate:

• Gate rising edge: Always fires one grain immediately.
• Gate held + DENSITY away from noon: Additional grains fire continuously at the density rate.
• Gate held + DENSITY at noon: Only the rising-edge grain plays; no additional grains while held.
• Gate low: No grains fire.

Use Gates mode when you want Particules to sound only while a gate is active — gating the granular texture from an envelope or a sequencer gate output.

Grain Parameters

Rememeber, parameters are read once when a grain starts and held for the grain’s entire duration. Changing a parameter mid-grain has no effect on that grain — it only affects the next grain seeded.

TIME

TIME controls which position in the recording buffer each grain reads from.

• Fully CCW: Grains read from the most recently recorded audio — close to live.
• Fully CW: Grains read from the oldest audio in the buffer — furthest from the record head.
• 12 o’clock: Midpoint of the buffer.

Patching a slow LFO into the TIME CV input makes the grain position drift back and forth through the buffer — a lush, swimming texture. A ramp LFO scans in one direction continuously, which is useful for time-stretching a frozen buffer (see Delay Mode and Techniques). A triangle LFO reads the buffer back and forth — scanning through older and newer material alternately.

PITCH

PITCH transposes grain playback from −24 to +24 semitones.

• 12 o’clock: No transposition (unity pitch).
• The knob has virtual notches at musically useful intervals — unison, perfect fifth (±7 semitones), octave (±12 semitones), and compound fifth (±19 semitones) — making it easier to land there when adjusting the knob.
• When the PITCH attenurandomizer is fully CW, the PITCH CV input tracks 1V/oct for standard keyboard or sequencer control.

SIZE

SIZE controls the duration of each grain and the direction of playback.

• Around 11 o’clock: Very short grain (approximately 30ms) — percussive, clicking texture.
• Turned CW from 11: Grain duration increases, up to about 4 seconds.
• Turned CCW past 11 o’clock: Grain plays in reverse, up to about 4 seconds in reverse. The further CCW, the longer the reversed grain.

SHAPE

SHAPE controls the amplitude envelope applied to each grain.

• Fully CCW: Rectangular (flat) envelope — the grain starts and ends abruptly. Percussive, clicking, good for rhythmic textures.
• 12 o’clock: Roughly symmetrical bell-shaped envelope — standard smooth grain.
• Fully CW: Slow attack, fast release — grain fades in slowly and cuts off. This creates a reversed, swell-like character even on forward-playing grains.

The amplitude envelope shape (SHAPE) is independent of playback direction — reversed grains still use the same envelope.

FREEZE

FREEZE stops recording to the buffer. The existing buffer contents are locked in place, and grains continue to play from the frozen material indefinitely.

• FREEZE button: Latching — press to freeze, press again to release.
• FREEZE CV input: A gate signal holds FREEZE for its duration.

Freezing is one of the most powerful tools in Particules. Freeze an interesting moment in your audio source, then sculpt the texture with TIME, DENSITY, PITCH, and the attenurandomizers — entirely disconnected from the live input.

Output

FEEDBACK

FEEDBACK controls how much of the output signal is mixed back into the input and fed through the granular processing chain again. It’s the feedback path around the entire granular engine.

• At low levels: subtle thickening as grains echo through the system.
• At higher levels: building, self-sustaining textures. The character of the limiting depends on quality mode — brickwall in Bright Digital, tape saturation in Scorched Cassette.
• Important: The feedback path routes back to the granular engine input, before the reverb stage. The reverb is not in the feedback loop.

In Wavetable mode, FEEDBACK changes function: it selects which of the 24 wavetable banks is played (see Wavetable Synth Mode).

DRY/WET

DRY/WET crossfades between the dry input signal and the processed (wet) granular output.

• Fully CCW: 100% dry — you hear the unprocessed input signal.
• Fully CW: 100% wet — you hear only the granular output.
• Center: Equal mix of both.

The reverb is applied after this crossfade and affects both the dry and wet signals.

In Wavetable mode, DRY/WET adjusts the balance between the continuous wavetable oscillator signal and the granularized version of it.

REVERB

REVERB adds reverb to the output. It applies to both the dry and wet signals — you hear reverb even if DRY/WET is fully CCW.

The tone and tail length of the reverb change with quality mode. In Bright Digital the reverb is tight and clear; in Scorched Cassette it’s long, warm, and slightly unstable.

CV inputs for FEEDBACK, DRY/WET, and REVERB

Each of the three output parameters has its own dedicated CV input jack, paired with an attenuverter knob. Patch a CV signal and use the attenuverter to set the modulation depth. The attenuverter is bipolar — fully CW is maximum positive modulation, fully CCW is maximum negative modulation, center is no modulation.

Grain trigger output

When Grain trigger on R output is enabled in the context menu, the OUT R jack outputs a trigger pulse on each grain event instead of audio. Use this to sync other modules to Particules’ grain rate — trigger envelopes, advance a sequencer, or clock other processors.

When this option is off (default), OUT R outputs normal stereo audio.

Attenurandomizers

Each of the four grain parameters — TIME, PITCH, SIZE, and SHAPE — has a smaller knob beneath it. These are the attenurandomizers. They’re dual-function controls: what they do depends on whether a CV cable is patched into the corresponding CV input.

Without CV patched: internal randomization

When no cable is patched into a parameter’s CV input, the attenurandomizer controls how much that parameter is randomized independently for each grain.

• 12 o’clock: No randomization — every grain uses the exact value set by the main knob.
• Turned CW: Uniform randomization — each grain picks a random value drawn evenly across a range centered on the knob position. A small CW turn gives subtle variations within a narrow range; further CW widens the range, up to sweeping the full parameter span at maximum.
• Turned CCW: Peaky randomization — each grain is more likely to pick a value near the knob setting, but occasionally jumps to an extreme. Further CCW means more dramatic and frequent jumps to the extremes.

This internal randomization is a good way to give granular textures an organic quality. A texture with no randomization can sound mechanical and repetitive; a texture with too much can sound chaotic.

With CV patched: attenuation and modulator control

When a CV cable is patched into a parameter’s CV input, the attenurandomizer changes function:

• Turned CW from 12: Standard attenuator — increases the depth of the incoming CV’s effect on the parameter. Fully CW = full CV depth.
• Turned CCW from 12: The incoming CV controls the depth of the internal randomization. Positive voltages increase how much the internal random source modulates the parameter. This is modulating your modulator — using an external CV signal as a VCA on the internal random source.

In Delay and Wavetable modes

The attenurandomizers work differently in Delay Mode and Wavetable Synth Mode — see those sections for details.

Context Menu

Right-click the Particules panel in VCV Rack, or select the Options in the Module menu on the MetaModule, to open the context menu. Special modes and configuration options are here.

SEED CV mode

Triggers (default) or Gates. Controls how the SEED CV input behaves. See Grain Generation for full details.

Delay mode

Activates Delay mode. See Delay Mode.

Wavetable mode

Activates Wavetable Synth mode. See Wavetable Synth Mode.

Lock pitch

Off (default), Octaves, or Octaves + 5ths. Rounds each grain’s pitch to the nearest locked interval.

• Off: Grain pitch set freely by PITCH knob and CV.
• Octaves: Each grain’s pitch is rounded to the nearest octave from the knob setting.
• Octaves + 5ths: Grain pitch rounded to the nearest octave or fifth.

Grain trigger on R output

When enabled, OUT R outputs a trigger on each grain event instead of audio. See Output.

Auto Gain

When enabled (default), Particules monitors the input signal and automatically adjusts input gain. Toggling this option on re-triggers the auto-gain calibration process.

Clear buffer

Clears the recording buffer immediately. Useful if you want to start fresh without repatching.

Manual Gain

Available only when Auto Gain is off. Sets a fixed gain applied to the input signal: 0 dB is unity (no change), positive values amplify, negative values attenuate. Range is −60 to +32 dB. The current value is shown to the right of the menu item.

Special Modes

Delay Mode

Activate Delay mode from the context menu. In Delay mode, Particules maintains a single permanent grain that continuously reads from the recording buffer — like a read head on a tape loop. The result is a true delay effect with extensive tonal shaping.

Delay time

Three parameters interact to set the delay time.

SIZE sets the delay buffer length — the ceiling for all delay times. At maximum SIZE, the full buffer is available. At lower SIZE values, the buffer shrinks, and all delay times shrink with it. If you’re not hearing the delay you expect, check SIZE first.

DENSITY sets the base delay time:

• 12 o’clock: Base delay equals the full buffer (the longest possible delay at the current SIZE).
• Turned away from 12: Base delay shortens exponentially, down to audio rates.
  • CCW: Single delay tap.
  • CW: Adds a second tap at a golden-ratio offset (multi-tap delay).

TIME extends the base delay set by DENSITY. At fully CCW, the delay equals DENSITY’s base value. Turning CW multiplies it exponentially, up to the SIZE ceiling. When DENSITY is at 12 o’clock, TIME has no effect — the base delay already fills the buffer. The further DENSITY is turned from noon, the wider TIME’s effective range.

Clocked delay

Patch a clock into SEED to sync the delay to an external tempo. The base delay time becomes the interval between clock pulses. DENSITY then selects a subdivision:

• CCW from 12: Binary subdivisions (1/2, 1/4, 1/8, 1/16).
• CW from 12: A wider range of rhythmic ratios.
• 12 o’clock: 1:1 — delay matches the clock interval.

You can also tap-tempo by patching a manual trigger source into SEED.

Pitch shifting in Delay mode

PITCH applies classic rotary-head pitch shifting to the delayed signal. At 12 o’clock, pitch shifting is bypassed. Shift up an octave or fifth and add FEEDBACK for shimmer-style effects.

SHAPE in Delay mode

SHAPE applies a tempo-synchronized amplitude envelope to the delay repeats. At fully CCW, no envelope is applied (standard delay behavior). Turning CW adds rhythmic gating or tremolo to the repeats.

Modulation in Delay mode

In Delay mode, the internal random sources for the attenurandomizers switch from stepped random to slow, smooth LFOs. This changes what the attenurandomizers do: instead of adding grain-to-grain variation, they apply slow continuous modulation to the delay signal. With no CV patched, the attenurandomizer knobs control the depth of that LFO modulation. With CV patched, they attenuate the incoming signal as normal — but any CCW-from-center portion routes the LFO as the modulation source rather than noise.

This makes chorus, vibrato, and flanger effects natural to set up: patch nothing into PITCH CV, turn the PITCH attenurandomizer slightly CW, and the delay pitch gently wobbles at LFO speed. Do the same with TIME for a slow, drifting echo.

Beat Slicer

Beat Slicer mode is accessed by engaging FREEZE while in Delay mode. A slice of the recording buffer is continuously looped and can be manipulated in real time.

• DENSITY: Sets the length of the looped slice (or selects a subdivision of an external clock).
• TIME: Selects which slice of the buffer is played.
• PITCH: Transposes the slice (−2 to +2 octaves).
• SHAPE: Applies a tempo-synchronized amplitude envelope to the loop.

Wavetable Synth Mode

Activate Wavetable mode from the context menu. In this mode, Particules stops using the audio input and instead plays back internally stored waveforms from the Plaits wavetable model — making it a standalone granular wavetable synthesizer.

Controls in Wavetable mode

• FEEDBACK: Selects which of the 24 wavetable banks is played (instead of its normal feedback function).
• DRY/WET: Adjusts the balance between the continuous wavetable oscillator signal and the granularized version.
• PITCH CV input: Always tracks 1V/oct regardless of the attenurandomizer position.
• PITCH attenurandomizer: Controls the amount of pitch randomization applied to individual grains.
• FREEZE: Halts grain envelope generation and stops new grains.
• DENSITY, SIZE, SHAPE, TIME: Function as in normal granular mode.

Quality mode works the same way as in normal granular mode — see Audio Input & Quality.

Techniques

Lush textures and sustained pads

The classic granular application. Freeze a sustained sound (a chord, a drone, a vocal sustain), then sculpt it with grain parameters.

Starting point: Quality mode Yellow (Sunny Tape) for warmth. DENSITY slightly CCW of center (moderate, regular grain rate). SIZE at 2–3 o’clock (medium-long grains). SHAPE at 12 o’clock. TIME attenurandomizer slightly CW (small random position variation). PITCH attenurandomizer slightly CW (subtle pitch shimmer). DRY/WET fully wet. REVERB to taste.

Freeze a chord or vocal sample. Use a slow LFO on TIME (through the attenurandomizer, attenuated) to gently swim through the buffer. Use a second slow LFO on PITCH CV (attenuated to a semitone or less of range) for shimmer. Increase DENSITY for a denser cloud; decrease for more spacious, separated grains.

Shimmer reverb

You need an external feedback path. Split the output, run it through an attenuator, and patch it back into IN R (or merge it with your dry signal into IN L using a mixer). Set PITCH to +12 (one octave up — use the virtual notch). Set DRY/WET fully wet. Set REVERB up. Grains pitch-shift up by an octave on each pass through the feedback loop, creating the rising shimmer effect. Tune PITCH to a fifth (+7 semitones) for a warmer shimmer that works with more input material.

Pitch shifting

For smooth pitch shifting, settings matter:

• Use longer grain sizes (SIZE at 2 o’clock or more).
• Use a smooth, symmetrical grain envelope (SHAPE near 12 o’clock).
• Keep DENSITY high (dense grain cloud smooths over transitions).
• Set PITCH attenurandomizer at 12 (no randomization — consistent pitch shift).
• DRY/WET fully wet.

Rhythmic and glitch effects

Use Beat Slicer mode (Delay mode + FREEZE) with an external clock for rhythmic buffer manipulation. Or, in normal granular mode, use triggered grains with a rhythmic sequence for controlled granular rhythms.

In normal granular mode, set DENSITY at 12 (no auto grain generation), confirm SEED CV mode is set to Triggers (the default), and patch a rhythmic trigger pattern into SEED. Each trigger fires one grain. Use a V/oct sequence into PITCH to add melody. SIZE controls the grain length — keep it short for percussive hits, longer for sustained notes that blur together rhythmically.

In Beat Slicer mode, clock SEED to your master BPM and set DENSITY to select a short subdivision (1/8 or 1/16 of the clock interval). This sets the loop length and keeps it tempo-synced. Then modulate TIME with a random stepped voltage — TIME selects which slice of the frozen buffer plays, so each new value jumps to a different position. Works especially well on drum loops, giving a spontaneous beat repeat effect.

Time stretching

Freeze audio, then use a ramp LFO to scan through the buffer with TIME.

Freeze a short drum loop or melodic phrase. Patch a slow ramp LFO into TIME CV. Set the ramp rate very slow — slower than the original material’s tempo. The grains now scan through the frozen buffer at a slower rate than real time, stretching the material. Increase DENSITY for better quality stretching (more grains fill in the gaps). Adjust SIZE to taste: smaller sizes give a more accurate stretch; larger sizes give a smoother, blurrier result.

Chordal washes

Patch a fast arpeggiated sequence cycling through 3–5 notes into PITCH CV. Set the PITCH attenurandomizer fully CW (full CV depth). Each new grain seeds at a slightly different moment and snapshots a different pitch from the arpeggio. The result is a polyphonic wash of the chord tones playing simultaneously, sustained for the grain duration. Use FREEZE to lock a stable source texture and let the pitch sequence do the harmonic work.

Cinematic drones and feedback loops

Feed the output back into the input using a cable split, attenuation, and a mixer or spare input. Set quality mode to Scorched Cassette for maximum warmth and saturation. Set FEEDBACK high. The granular engine feeds into itself — grains spawning new material that spawns more grains. Add a filter in the feedback path to shape which frequencies accumulate. This creates long, evolving drones that are self-generating and respond to small parameter changes dramatically.

Karplus-Strong string synthesis

See Delay Mode for setup. The key insight: Particules’ delay engine at audio rate is a resonant comb filter — excite it with a short impulse and it rings at the delay frequency. FEEDBACK controls how long it rings (sustain). DENSITY CV at 1V/oct controls pitch.

Particules as a multi-FX unit

Particules can process external audio (guitar, synth, hardware) simultaneously with pitch shifting, granular echo, and reverb. Set DRY/WET to blend the original with the processed signal. Use a gentle grain density (slightly CCW) for a subtle echo effect rather than a full granular texture. Quality mode Red (Scorched Cassette) adds wow and flutter to the processed signal.

Run an external instrument into IN L. Set DRY/WET to about 10 o’clock (mostly dry, some wet). Set DENSITY slightly CCW (sparse, regular grains). Set REVERB to taste. The result is a subtle granular echo + reverb effect — the instrument is still clearly audible but surrounded by a halo of processed material.