Beautiful Bleep

Voice Cells: polyphonic synths

A Voice Cell turns one synthesizer voice into a polyphonic instrument. Inside, you patch one voice, like a mono synth. The cell plays it once per note, up to 16 voices at a time.

Make one

  • Add → Cell → New Voice Cell gives you a cell with a starter voice inside: Oscillator → VCA, with an ADSR on the VCA's gain. It makes sound right away.
  • Add → Cell → Synthesizers → Basic Poly is a ready-made synth: Oscillator → Multi Filter (Rough) → VCA, one ADSR for the filter and one for the VCA.
  • Right-click an ordinary cell → Make Voice Cell turns it into one.

Connect it

From outside, a Voice Cell has:

Jack What
Notes all notes, e.g. from MIDI Device Notes
Out L / Out R all voices, mixed
your own inputs the same signal for every voice: an LFO, a cutoff macro, pitch bend
your own outputs summed over the voices

To play it from a sequencer or any pitch + gate source, put CV to Notes in between.

Design the voice

Double-click the cell. Inside are two fixed nodes:

  • Voice In gives this voice its note: Pitch, Gate, Trigger, Velocity, Release, Pressure, Voice # (0, 1, 2 …) and Random (a new value per note, for a little variation between voices).
  • Voice Out takes the voice's audio (Out L, Out R) and Active: patch the amp ADSR's Active here, so the voice is freed exactly when its sound has faded.

A knob you turn inside changes it in every voice. Add your own shared inputs and outputs by dropping a cable on Cell In's or Cell Out's "new" row.

Pitch bend: patch MIDI Device Bend into a shared input and, inside, into the oscillator's Pitch with amount 2/12 for ±2 semitones (that's how Basic Poly does it).

Settings

Click the cell to open its view:

Setting What it does
Voices 1–16
Allocation which free voice a new note gets: round robin, the lowest free one, or the voice that last played the same note
Stealing when all voices are busy: take the oldest, quietest, lowest or highest note — or none (the new note is dropped)
Mode Poly; Mono (one voice, every note restarts the envelopes); Legato (one voice, overlapping notes don't restart them)
Glide time to slide from one pitch to the next; Glide Mode: always, or only when notes overlap
Stereo Spread spreads the voices from left to right
Output plain sum, or compensated (quieter per voice, so 8 voices don't clip)

The dots on the node light up for the voices that play. Voices that don't play use no CPU.

Save it

Right-click → Save as preset, give it a name and a category (e.g. Synthesizers). It then appears under Add → Cell → in every project.