playgrid.ccRead tutorial

Cirrus Window — complete SuperCollider settings

Signal flow: shared pulse-width motion → three stereo-weighted saw/pulse voices → amplitude and serial bandpass/lowpass → stereo audio bus → three-tap ensemble → stereo room → bass-side filtering and DC removal → output gain.

Oscillators and envelopes

Control Value and audible role
Saw / pulse levels −13 / −16 dB per voice
Voice detune −7 / 0 / +7 cents, for sustained beating
Pulse detune Additional +5 cents relative to its saw
Pitch drift ±2.2 cents at 0.17 + absolute voice cents × 0.008 Hz
Drift phase 0.08 / 0.41 / 0.73 cycles, converted to radians
Pulse-width motion Shared 0.09 Hz sine, width 0.29 ± 0.025
Stereo voice weights L: 0.87 / 0.5 / 0.13; R reversed; sums divided by 1.5
Amplitude envelope Levels 0 → 1 → −2 dB → 0; times 0.85 / 2.5 / 2.4 s; sustain at the third level until gate release
Envelope shaping Power 1.6; effective sustain −3.2 dB
Voice gain −8 dB × 0.65 × amplitude envelope
Velocity sensitivity Gain multiplier 0.55 + 0.45 × clipped velocity
Filter envelope Levels 0 → 1 → 0.6 → 0; times 1.8 / 3 / 2.4 s, with gate-controlled release

Serial filters

Control Value and audible role
Bandpass BPF, base center 850 Hz, reciprocal Q 1 / 0.85
Envelope depth +0.8 octaves at peak
Key tracking +0.15 cutoff octaves per played octave, relative to MIDI 60
Lowpass LPF, 6200 Hz, following the bandpass

BPF uses bandwidth divided by center frequency as its reciprocal-Q argument. The broad setting focuses the midrange without a narrow resonant peak. Its response and the oscillator spectra differ from both Wavetable's filters and Csound's implementation.

Ensemble

Control Value and audible role
Input highpass HPF, 220 Hz
Delay units Three DelayC taps, 40 ms maximum allocation
Center / depth 20 ms / ±3.6 ms
Rate / phase 0.24 Hz; 0, 2π/3, 4π/3 radians
Tap inputs Left, stereo mean, right
Wet channel weights 75% outer tap + 25% center tap
Dry / wet 78% / 22%
Side gain 1.05 after the ensemble blend

Stereo room

Control Value and audible role
Predelay DelayN, 28 ms, on the room input only
Reverb Stock FreeVerb2, two inputs and two outputs
Internal mix 1, wet output only
Room / damping 0.84 / 0.55
External dry / wet 75% unpredelayed ensemble / 25% room

FreeVerb2 replaces the room algorithm; it does not implement Ableton's room or Csound's reverbsc. Its room and damping values are dimensionless, without a direct conversion to seconds or a cutoff in hertz. The wet blend and predelay preserve the same broad sound-shaping intent.

Output and voice behavior

Control Value and audible role
Bass-side filtering HPF at 180 Hz on the side channel only
DC removal LeakDC, default coefficient 0.995
Finishing gain +10.9 dB, stage 5 only
Fixed gain 0.5 after the stage selector, all stages
Stage selector ~soundStage = 5, range 1–5; evaluate the block again to replace the active instrument
Note arguments ~soundNote.(midi, velocity, hold); defaults 60, 0.79, 2.5 s
Release behavior Gate closure releases amplitude and filter contours; the voice frees after its amplitude release
Processor order Shared motion Synth → voice group → shared effects Synth

Saw and Pulse have their own startup phases and band-limiting behavior; the phase settings above apply to pitch drift, not oscillator resets. The source wavetable morph and Shimmer are interpreted as a saw/pulse blend with slow pitch drift. No external sample, plugin or SynthDef is required. The connected full SuperCollider script ends the build guide.

Unit references

SuperCollider BPF explains reciprocal Q. SuperCollider FreeVerb2 documents the room and damping controls.