VX-7

A preset-based six-operator FM voice synthesizer — a complete playable instrument, not a bare oscillator. Sixteen voices live in its bank, from electric pianos and basses to bells, pads, and glassy textures; pick one with the Preset knob (or sequence them per-note over the PATCH input) and play it from any pitch + gate pair. Notes overlap by their release tails, so a fast line with long releases blooms into a chord-like wash of up to eight voices.

How it works

Each voice is six sine operators wired by one of the 32 classic DX-7 algorithms: carriers you hear, modulators that bend the carriers' frequency to enrich the timbre — classic DX-style FM. The algorithm card shows the routing live (dark boxes are carriers, white are modulators — click one to edit that operator); each algorithm fixes which operator carries the feedback loop, and the Feedback knob sets how hard it bites. Every operator has its own 4-rate/4-level envelope (the authentic DX contour): a note travels level 1 → 2 → 3 and holds at level 3 until you let go, then falls to level 4 — so a voice can strike bright, settle fast, and fade slowly all in one gesture (that's the electric pianos and bells). Each sounding voice also carries its own resonant low-pass filter (with its own ADSR), a full LFO (six waveforms, onset delay, pitch and amplitude depth, key sync), onset pitch transient, and portamento. A voice locks onto the selected preset at note-on and keeps it for its whole life — switching presets never disturbs a ringing tail. Every parameter of every slot is editable on the panel: pick a slot and an operator, and shape its frequency, level, envelope, routing, LFO, and filter directly; some slots are mono (single-voice legato — Sync Lead, Sub Bass), the rest polyphonic.

Controls

  • Preset — the active voice slot, E.Piano through Init (16 slots).

Overridden by the PATCH input while it's connected.

  • Volume — master level, 0–100%. The default sits a single voice well

inside ±5 V; big stacks compress gently instead of clipping.

  • Bright — global FM depth, 0–2×. 0 strips every voice to pure sines,

1× is the preset's authored tone, 2× pushes it aggressive.

  • Time — scales every envelope duration at once, 0.25–4×. Below 1×

everything snaps; above, attacks and tails stretch. Sustained levels don't change — only how long it takes to get there and away.

  • Glide — portamento. 0% jumps instantly; higher settings slide from the

previous note's pitch, up to a second or two across wide intervals.

  • Punch — an onset pitch transient, ±12 st. Positive starts each note

sharp and falls to true pitch in ~55 ms (scaled by Time) for a percussive thump; negative scoops up into the note. 0 is off.

  • Cutoff / Reso — center-detented sweeps riding on top of the preset's own

filter. At center you hear the preset unchanged; turn down toward fully closed/neutral, up toward wide open / singing self-oscillation — the whole travel works from any preset.

  • Env — scales the preset's baked filter-envelope motion, 0–100%. At 0

the filter sits still no matter what the preset authored.

Inputs & outputs

  • V/OCT — note pitch, 1 V/oct, 0 V = middle C. A held note tracks this

live (through Glide), so you can bend it under a held gate.

  • GATE — rising edge starts a note, falling edge releases it. Detection is

hysteretic (fires at ~1 V, re-arms below ~0.1 V), so slow or wobbly gates don't double-trigger.

  • PATCH — live slot selector, 0–10 V across the 16 slots. While

connected it overrides the Preset knob; because voices latch their preset at note-on, sequencing it plays a different instrument per note while earlier tails keep ringing with theirs.

  • L / R — the stereo voice mix (voices sit center; the sum is mono-safe).

Tips

  • Sequence PATCH from a step sequencer alongside pitch and every step can be a

different instrument — tails from the previous instrument hang over the new one.

  • Bell or Glass with long releases + a fast arpeggio is an instant

shimmering wash; add negative Punch for soft mallet blips.

  • Turn Bright down to 0× on any preset for its pure-sine skeleton — useful

for sub-layers under another voice.