Processor Errant

s3g Processor Errant is a zero-input, stereo, eight-voice heavy bass synthesizer that composes repeatable glitch genealogy around a playable analog-style core. A band-limited saw/pulse/triangle body, protected sub oscillator, driven four-stage resonant low-pass filter, and snappy pitch and filter contours keep the MIDI note clear enough for growling basslines, sub-bass, gabber, and hard techno. Errant captures that performance into a causal family archive and lets later notes inherit it through recurrence, reversal, rate changes, ratchets, prediction error, block loss and replacement, quantized reconstruction, polarity edits, folding, and stereo exchange. Those processes disturb the upper body while the oscillator spine and sub remain intact. No source samples are distributed with or loaded by the instrument.

Errant sits near Processor Fault and Ambi Encoder Stochastic in the instrument family, but it gives their unstable processes a different performance grammar. Fault continuously interprets bytes, waveforms, codecs, and transmission processes. Stochastic evolves breakpoint-waveform populations. Errant makes the MIDI sequence itself a mutation program: repeated notes recall a lineage, intervals transform it, and simultaneous notes branch as siblings from the same archive snapshot.

Errant keeps its mutation grammar and protected bass circuit on one performance surface, with direct triggering available for auditioning without a MIDI controller.

Workflow

  1. Insert s3g Processor Errant on a stereo instrument track and play MIDI notes, or press GENERATE in the editor to audition middle C.
  2. Route its ordinary stereo output to the main stereo speakers or the project's established monitor path.
  3. Choose CELL, PHRASE, or FIELD for the intended time scale.
  4. Raise GROWL from a rounded analog body into pressured harmonics and formant bite, then place the voice with REGISTER, KEY ROLE, and PITCH GRAVITY.
  5. Build the bass circuit with SUB, CUTOFF, RESONANCE, CONTOUR, and DRIVE. Add CROSSWIRE after the stable body is established.
  6. Set ANCESTRY for continuity with earlier events and MUTATION for the severity of each transformation.
  7. Use SPAN, DENSITY, and REPEAT to establish the phrase scale and recurrence.
  8. Select a stereo topology, then set COHERENCE and WIDTH. The low body remains centered by design.
  9. Change SEED to create another repeatable genealogy. The same seed, parameters, and MIDI performance reproduce the same result.

Event Grammar

ControlMeaning
MODECELL makes a finite one-shot, PHRASE makes a longer self-mutating utterance, and FIELD continues while held before following its release.
SPANMoves from compact event cells toward longer segments and phrase memory.
DENSITYIncreases the admission of impulses, cuts, and active descendants.
ANCESTRYControls how strongly a segment reuses audible history from the current note or the shared family archive instead of returning to fresh synthesis.
MUTATIONRaises the probability and depth of reverse reads, held rates, quantized levels, prediction errors, block failures, replacement, folding, ratchets, polarity changes, and related edits. Discontinuous operations enter through short transitions rather than raw sample jumps.
REPEATRaises recurrence and makes a repeated MIDI note more likely to recall an established family fragment.
COHERENCEMoves the two lanes from related but independent timing and pitch behavior toward shared decisions.

The phrase memory is causal: a voice can only transform sound the instrument has already generated. Every sounding voice contributes to a four-second family archive. A new note takes a fixed snapshot of the preceding archive, so notes arriving together receive the same ancestor and become siblings rather than feeding instantaneously into one another. The new voice also owns a local memory whose descendants are written back alongside some fresh source. Long notes therefore contribute more history, and later notes accumulate performance history instead of selecting unrelated random effects.

Note Relationships

Mutant Bass Circuit

GROWL pressures a continuous upper body derived from the same oscillators as the fundamental. It no longer selects a replacement sound source, so high settings cannot crossfade the bass entirely into noise:

REGISTER offsets the center by ±36 semitones. KEY ROLE determines whether MIDI note number addresses ancestor and inherited-fragment pitch, event clocks, or both. PITCH GRAVITY controls the strength of equal-tempered tracking and how firmly the stochastic period walk is pulled back toward the played note. Velocity sensitivity determines how much note velocity controls final excitation level; velocity always remains part of the mutation relationship.

ControlBass-circuit behavior
SUBAdds a band-limited pulse/sine voice one octave below the played oscillator. A restrained protected path remains centered after the main filter and drive stages.
CUTOFFSets the nominal four-stage low-pass frequency from approximately 45 Hz to 9.9 kHz before note tracking and modulation.
RESONANCERaises bounded filter feedback. Saturation occurs inside the cascade so high settings reinforce rather than detach from the bass body.
CONTOURMoves cutoff over the note: positive values begin open and close, while negative values begin closed and open toward the nominal cutoff.
DRIVERaises input pressure and stage saturation inside the filter, followed by restrained output saturation.
CROSSWIRERoutes a smoothed, segment-level fault voltage into upper-oscillator detune, pulse width, cutoff, resonance, wire strands, and genealogy level. It is intentionally unstable but cannot displace the MIDI-rooted oscillator spine.

Glitch operations are treated as damaged circuitry rather than accidental DSP failure. The event genealogy is now an upper-body mutation layer around a continuous oscillator spine. Segment boundaries and memory-loop wraps use extended crossfades; ratchets use smooth windows; dropped or replaced blocks slew between states; held-rate and quantized samples are de-stepped; and Crosswire, filter, voice-count normalization, voice stealing, output gain, and final sample motion are independently ramped. The pulse family uses a predominantly rounded analog pulse with only a restrained band-limited hard edge, and the throat bank is excited by continuous oscillator energy rather than an impulse train. The result can still fracture abruptly, but an isolated click should belong to the synthesized event rather than to an unsafe sample jump.

Stereo Topology

TopologyEvent relationship
SPINEPlaces related material near a shared center.
WINGSOpens related but decorrelated descendants to either side.
EXCHANGEAlternates event emphasis between left and right.
SIDEUses complementary polarity for deliberately lateral fragments.

WIDTH sets the strength of the selected relationship. A low-band split below approximately 125 Hz maintains a centered anchor even when descendants become strongly lateral.

MIDI and State

The instrument accepts CLAP note events and ordinary MIDI note-on and note-off messages, with a selectable omni or single-channel receive filter. MIDI CC 120 and CC 123 release all active voices. Eight voices are available; when all are occupied, the oldest is replaced. CELL behaves as a one-shot, PHRASE runs a finite genealogy that note-off may close early, and FIELD evolves until its note is released.

SEED is stored with every host parameter, including the complete bass circuit, MIDI receive, topology, and output gain. The momentary TRIGGER action is intentionally excluded from saved state so loading a project cannot generate a surprise note. Earlier project states migrate to the current bass circuit with safe defaults. Audio is DC-blocked and passes through a final bounded knee.