Node Reference & User Guide

1. Set the global controls. At the top of the page, set your base tempo in bpm, select a beat unit, and set the window length in seconds. A typical starting point is ♩ = 52, ♩=4, 60 sec.

2. Work on the tracks. The left sidebar lists your tracks. Click + add track to create a new layer. Each track is given a distinct color from the Preplan palette (ochre, teal, lavender, sage, rose, golden) and an auto-generated name like "Voice A". Rename a track by clicking into its name field.

3. Author segments. Inside each track, each card represents one segment with three fields: count × num / den. Change numbers directly, or pick the denominator from the dropdown (1, 2, 4, 8, 16, 32). Reorder segments with the up/down arrows, remove with the trash button, and add more with + add segment. The readout below the card shows the bar duration in milliseconds and the segment's cumulative contribution to the loop.

4. Read the timeline. The right pane shows every track as a horizontal lane. Downbeat ticks appear in the track's color; loop boundaries are tall ticks labeled with their time signature (e.g. 5/8). Hover over any tick to see a readout below: track name, bar number, time signature, absolute start time, and bar duration in milliseconds.

5. Inspect the loop periods. Scroll down below the timeline to see a card for each track listing its period length and how many times the loop fits in the window. This is where you diagnose whether your layers settle into a long common cycle or drift continuously.

The simplest polymetric pattern. Voice A plays twelve bars of 3/4; Voice B plays twelve bars of 4/4 at the same tempo. Voice A's loop completes in 36 quarter-note beats; Voice B's loop completes in 48. The two lines realign every 144 beats (the LCM).

In the timeline you will see Voice A's downbeats land three-for-every-four of Voice B's. Coincidences appear at every bar of Voice B that also happens to be a bar of Voice A — which, for this combination, is every fourth bar of A and every third bar of B. Set the window to 60 seconds and watch the coincidence density thin out as the layers drift.

This is a useful warm-up sketch: it tells you how long a full polyrhythmic cycle actually takes before you commit to any pitches.

Two tracks with complementary irrational meters. Voice A's loop is 35 eighth-notes long; Voice B's loop is also 35 eighth-notes long — so the two layers realign exactly at each full loop. Between realignments they drift through an evolving sequence of offsets.

Set the beat unit to ♩=8 so the eighth-note gets the tempo beat. You should see a coincidence at every loop boundary plus a few internal ones where the 7-grouping and 5-grouping briefly line up. This is the engine behind a lot of Balkan-tinged and post-minimal writing.

To make it breathe, add a third track with an unrelated length — try Voice C: 3x11/8. The three-way realignment cycle is now much longer, and the coincidence map becomes a compositional feature in itself.

This is the default sketch that loads with Preplan. Each track is itself a meter sequence — not a single repeating bar but a through-composed cycle that restarts after all segments have played. The two voices have different total durations so they gradually shift against each other every time one loop completes before the other.

Look at the loop-start markers (tall ticks with time signatures) — you will see Voice A announce 4/4, then 7/8 four bars in, then 5/8 seven bars in, then restart with 4/4. Voice B does the same with its own segment sequence. The two tracks slide in and out of alignment across the whole window.

This is the architecture Ligeti used in San Francisco Polyphony and the first book of Pétudes — independent metric lines that share a pulse but never settle into a common bar. Use send to Compose to turn the sketch into a movement, then develop each voice's pitch content independently in the Graph view.

A single-track sketch used to plan an accelerating surface. The base pulse stays the same (♩=60) but the bar length shrinks progressively: four bars of 4/4, then four of 3/4, then 2/4, then 1/4. Downbeat density doubles by the end of the loop.

Use the readout to see bar durations drop from 4000 ms (4/4) to 1000 ms (1/4). This is a compositional strategy for notating a gradual acceleration as a meter ramp instead of a tempo change — players count the same beat throughout, but the music gets denser because bars arrive more frequently.

Mirror the ramp on a second track with the reverse sequence (4x1/4, 4x2/4, 4x3/4, 4x4/4) to create a converging-then-diverging density pattern across two voices — a useful texture for duo passages.

A minimalist-style pair where both tracks reuse the pair of numbers {7, 11} but in swapped positions. Voice A's loop is 7 × 11 = 77 eighth-notes; Voice B's loop is 11 × 7 = 77 eighth-notes. They realign every 77 eighths — roughly every 24 seconds at this tempo — but the barline patterns inside that cycle are completely different.

The coincidence dots at the bottom of the timeline will be sparse in the middle of each cycle and clustered at the loop boundaries. This is the effect Reich and Adams exploit when two lines share an overall period but differ in their internal bar groupings.

Once the sketch feels right, send to Compose. Add a third voice in a related meter (say Voice C: 5x7/8, 5x11/8) to introduce a drifting contramotion on top of the realigning pair.

1. Pick the EDO. Use the EDO dropdown in the header. Switching away from a higher EDO drops steps that no longer exist (e.g. going from 31 → 12 keeps only steps 0–11).

2. Toggle steps. Tap any dot on the ring to select / deselect it. Hit targets are generous (14 px) — pen and finger friendly. Selected dots fill in with the progression's accent colour; the polygon between them redraws automatically.

3. Rotate. The two circular arrow buttons below the circle rotate the pattern by −1 / +1 step. Great for auditioning modal rotations of the same shape.

4. Name, duplicate, delete. The text field below the circle is the circle's name (used in playback and when sent to Compose). Hover the card to reveal duplicate and delete icons.

5. Add more circles. The + Add circle placeholder at the right of the strip inserts a new empty circle at the end of the sequence.

Start with a Harmonic Series node set to your chosen fundamental (e.g., 131 Hz for C3). Generate 16 or more partials to create a rich harmonic field. Connect this to one input of a Spectral Morph node, and connect a second Harmonic Series (with a different fundamental) to the other input. The morph parameter (0.0–1.0) controls continuous interpolation between the two harmonic worlds — automate it with a BPF for gradual harmonic transformation over time.

Feed the morphed output through a Formant Filter to impose vowel-like resonance characteristics on the pitch material. Set formant centers to approximate vowel frequencies (e.g., [270, 2300, 3000] for an "a" vowel, [400, 2000, 2600] for an "e" vowel). This creates a timbral connection between the abstract pitch material and the human voice, a hallmark of spectral composition.

Connect the filtered pitches to a Chord node, pair it with a Rhythm Tree, and assemble everything in a Voice node for notation export. For additional refinement, use the Noise-Tone Axis node to map a BPF to articulation markings that trace the tone-to-noise continuum across the passage.

Begin with a Tone Row node and enter your twelve pitch-classes (0–11). The node automatically generates all four canonical forms: Prime (P), Retrograde (R), Inversion (I), and Retrograde-Inversion (RI). Each form appears as a separate output port.

Connect any form to a PC Transpose node to produce transposed versions (P0, P1, P2... through P11). Use Chain nodes to concatenate different row forms into longer pitch sequences. For integral serialism, pair the row with a Duration Series node that derives rhythms from the same or a related numeric series, ensuring that pitch, rhythm, and dynamics are all governed by unified serial logic.

For Webern-style pointillistic textures, spread the row across wide registral leaps using Interval Expand with a large factor, then assign sparse rhythms with many rests using Insert Rests. For denser Schoenberg-style writing, keep the row in a compact register and use flowing rhythm trees.

Feed a cantus firmus (a Pitch List or any pitch source) into a Counter-line node. Start with Species 1 (note-vs-note) and the default dissonance at 0.2 — Fux-strict territory. The output counter port is a complete counter melody conforming to your scale. For modal CP, pick a Dorian or Phrygian preset; for modernist writing, push the dissonance slider to 0.5+ and pick the Hindemith style preset.

Wire the Counter-line's counter output and the original cantus into a CP Check node as upper and lower voices. The 0–1 score grades strictness; the breakdown output lists violations by category (parallel 5ths, unresolved dissonances, etc.). Use CP Check's tension output — a per-pair 0–5 curve — as a BPF source to drive dynamics that breathe with your voice-leading density.

For canon, route the Counter-line into an Imitation node. Pick Transform P (prime) and Interval 7 for Bach-style canon at the fifth, with Delay 2. The node emits dux, comes, and pre-built rhythm trees with the delay baked in as leading/trailing rests. Pair each voice with its rhythm in a Voice node and stack them in Voice Layer for a complete playable canon. Try transform I for inversion canon, R for crab canon, or RI for Webern-style retrograde-inversion.

For invertible (double) counterpoint, feed any working two-voice pair into a Double CP node. Pick the swap interval — 12 (at the octave), 16 (at the 10th), or 19 (at the 12th, Bach's favourite). The editor shows stacked original and inverted rolls with ⚠ flags on every pair that breaks after the swap, so you can see at a glance whether the passage is truly invertible or needs rewriting.

Start with a short repeating pattern in a Pitch List — typically 8–12 notes. For Clapping Music, encode rests as negative midicent values inside the pattern (e.g. -1): cp.phase treats any non-positive entry as a rest slot. For melodic phasing (Piano Phase, Violin Phase), use 8–12 pitches from the piece's scale.

Feed the pattern into a Phase node. Set cycles to how many pattern repeats you want — 12 is the classic one-traversal length, 48 gives a ~5-minute extended realisation. Set voice count to 2 (for Clapping / Piano Phase), 3 (for Drumming), or 4 for a richer cascade. Leave shift per cycle at 1 for Reich-authentic drift. Voice 0 anchors; voice i rotates by i × shift notes per cycle. When cycles × shift is a multiple of the pattern length, the voices re-align at the end — the editor lights up a "voices re-align at end" indicator when this is true.

Each voice output pairs with its matching rhythm output (voice1 + rhythm1, voice2 + rhythm2, etc.). The rhythm trees are pre-padded to identical total lengths, so wiring each (pitch, rhythm) pair into its own Voice node and stacking everything in a Voice Layer produces a playable phased texture with no alignment math. For Clapping Music specifically, both voices use the same pitch (say, A4) — the music is the phase relationship. For Piano Phase, transpose one voice up an octave via the Voice node's clef or add a Transpose node before the Voice.

Define two pitch-class sets using PC Set nodes. Set A might be a trichord like [0, 1, 4] and Set B another trichord like [0, 3, 7]. Connect both to a Pitch Multiply node. The multiplication works by transposing each pitch of Set A by every interval derived from Set B, then taking the union of all resulting pitches.

The result is a larger pitch field that contains the intervallic DNA of both source sets. Run the output through Sort to arrange pitches registrally, or through Unique to remove duplicates. This technique, developed by Boulez for works of the 1950s and beyond, produces harmonically rich fields from small generating cells — the compositional equivalent of multiplying two numbers to get a product that contains both factors.

Experiment with different set combinations: multiply a whole-tone fragment by a chromatic cluster, or multiply two segments of a tone row. The results vary dramatically depending on the intervallic content of the source sets.

Build a deeply nested Rhythm Tree with multiple layers of tuplet subdivision — for example, a quintuplet where individual beats are further divided into septuplets or nested triplets. This creates the characteristically dense notational surface of new-complexity writing.

Feed the rhythm tree through Insert Rests with a binary mask (e.g., [1,0,1,1,0,1,0,0,1,1]) to selectively silence events. The proportional structure is preserved but certain attacks are removed, creating rhythmic patterns of presence and absence that break up the regularity of the underlying subdivisions.

Then apply Rest Density with a BPF curve to add an additional layer of time-varying rest probability — perhaps beginning dense and gradually thinning out, or the reverse. Finally, connect Grace Notes to add Ferneyhough-style gestural ornament clouds around the surviving main pitches, with density and range parameters controlling the proliferation of ornamental material. The result is a richly layered rhythmic surface where structural rhythm, rest patterning, and gestural ornamentation interact.

Start with stochastic raw material: Random Pitches for uniformly distributed pitches across a range, or Random Walk for a melodic line that wanders organically with constrained step sizes. Set the range and quantization to match your desired ambitus and temperament.

Filter the result through a Sieve node with modular congruence conditions. For example, moduli [3, 5] with residues [0, 1] retains only pitches at indices that are either divisible by 3 or have remainder 1 when divided by 5. This produces the irregular, non-periodic pitch distributions characteristic of Xenakis's sieve-based compositions — asymmetric scales and pitch selections that sound neither tonal nor uniformly chromatic.

Optionally pass the sieved material through a Permute node with a specific index sequence to impose a particular ordering. Combine with stochastic rhythms for fully Xenakis-inspired textures, or pair with structured rhythm trees for a hybrid approach.

The key insight: use the same BPF curve to drive both pitch contour and rhythmic density simultaneously. Create a BPF node with a shape that describes the energy profile of your passage — for example, a slow ascent followed by a rapid descent.

Connect this BPF to an Apply Profile node to shape your pitch material: where the curve is high, pitches are pushed to the upper register; where low, to the lower register. Simultaneously connect the same BPF to a Density Profile node: where the curve is high, rhythmic events are densely packed; where low, they are sparse.

The result is tight spectral-temporal integration — high pitches coincide with dense rhythms, low pitches with sparse rhythms (or vice versa if you invert one of the curves). This creates a unified gestural quality where register, density, and energy move together as a single compositional parameter, a technique used extensively in spectral and post-spectral composition.

Use a Logistic Map node (r between 3.57 and 4.0 for chaotic behavior) or a Henon Map to generate deterministic chaotic sequences. These are not random — they are fully determined by the initial parameters — but they produce sequences that are effectively unpredictable and never repeat.

The output values (typically 0.0–1.0 for Logistic Map) need to be mapped to musical ranges. Connect to a Map Range node to convert to midicents (e.g., map [0, 1] to [4800, 8400] for a 3-octave range). Then pass through Microtonal Quantize to snap to your desired temperament (12 for semitones, 24 for quarter-tones).

The Henon Map is especially powerful because it outputs two correlated sequences (X and Y). Map X to pitches and Y to rhythm durations (via Map Range and Quantize Rhythm) so that both dimensions of the music emerge from the same chaotic process. Small changes to the initial parameters (r, a, b, x0, y0) produce radically different musical results, making these nodes excellent for generating variations.

Use waveform generators to create smooth, periodic dynamic contours. A Gen Sine node produces undulating dynamics that cycle between loud and soft. Gen Triangle gives linear ramp-up and ramp-down patterns. Gen Sawtooth creates repeated crescendo-to-subito-piano gestures. Gen Square alternates abruptly between two dynamic levels.

Connect the waveform output to a Velocity Curve node, which samples the BPF at the number of events in your rhythm and maps the values to MIDI velocities (0–127). For simpler cases, the Cresc/Dim node provides a direct crescendo or diminuendo with adjustable curve shape — linear, logarithmic, or exponential.

For complex dynamic shapes, combine Gen Fourier (additive synthesis of control curves) with Velocity Curve. Or draw a completely free BPF by hand and sample it. Connect the resulting velocity list to the velocity input of the Voice node to apply the dynamics to your composition.

Two approaches to microtonal quantization. The first uses Microtonal Quantize with any equal division of the octave: 24 for quarter-tones (the most common microtonal system in contemporary Western notation), 36 for sixth-tones, 48 for eighth-tones, or any other integer. This snaps free-pitched material to a regular microtonal grid.

The second approach uses Spectral Round, which snaps each pitch to the nearest harmonic partial of a given fundamental. This produces just-intonation-like tunings derived from the harmonic series — the pitches are irregular (not equally spaced) but acoustically coherent. Combine with Harmonic Series and Formant Filter for a fully spectral approach to microtonality.

For custom tuning systems (Turkish makam, gamelan slendro, just intonation scales, etc.), use Scale Quantize with a list of scale degrees defined in midicents offsets within the octave. Any tuning system that can be described as a set of pitches within one octave can be used as a quantization target.

For large-scale compositions, build each section as a separate sub-graph with its own source-transform-voice chain. Each sub-graph produces a complete Voice output. Name each section with a unique ID using the Section node.

At the top level of your composition, use Section reference nodes to point to your pre-composed sections. Feed these into a Voice Chain node (which accepts up to 8 inputs) to arrange them in temporal order. An A-B-A form, a rondo, or a through-composed sequence can all be assembled this way. You can reference the same section multiple times, with different transformations applied each time (e.g., transposed, with different dynamics, or with modified rhythms).

For polyphonic textures, use Voice Layer to stack multiple voices simultaneously, then chain layered sections using Voice Chain for the full temporal structure of a multi-voice piece. This hierarchical approach — building small cells, combining them into sections, and assembling sections into forms — mirrors the way many composers think about large-scale structure.

The opening of Beethoven’s Fifth Symphony is built entirely from one four-note rhythmic cell — eighth-rest, three short notes, one long note — stamped at different pitch levels. That maps directly onto Grafer’s Pitch List + Rhythm Repeat + Voice chain. Enter the fate rhythm as proportional weights [−1, 1, 1, 1, 4] (the negative weight is the pickup rest, the final 4 is the held note), and set Measures to however many motif statements you want.

For the exact opening (bars 1–4), feed the Pitch List node the midicent sequence [6700, 6700, 6700, 6300, 6500, 6500, 6500, 6200] — that is G4, G4, G4, E♭4, F4, F4, F4, D4, the two famous fermata statements. In 4/4 time at 𝅝=108, two Rhythm Repeat measures cover them both. The Voice node maps each pitch to the next non-rest leaf of the rhythm tree, so the one rhythmic cell is silently repeated for every motif.

For the development that follows, feed eight motif statements of pitches into the same rhythm with Measures = 8. Each statement keeps the fate rhythm but drops its starting triplet by a step of the C-minor scale, with the long note landing a minor third below — so the motif cascades F→E♭, E♭→D, D→C, and so on, mirroring Beethoven’s sequential descent. The same technique works for any motivic piece: pick a distinctive rhythmic cell, supply a stream of pitches, and let the Voice node pour the pitches through the rhythm.