GRIND path

Four stages in fixed order: oscillator and unison, filter, shaper, bus.

Osc and unison

WT Pos morphs through the factory wavetable’s four frames: saw, folded saw, square, triangle. The readout names the nearest frame.
Every wavetable frame has its fundamental removed. The grind path contributes harmonics and texture; the fundamental is the sub’s job. This is why the two paths can be mixed freely without the low end changing level.
Level (up to +12 dB): the grind path’s own level. Fully down is silence, so the sub can be heard alone; the sub’s Level is stated against this one.
Voices (1 to 7): unison clones of the whole grind path, detuned in cents and spread across the stereo field. The offsets are fixed rather than random, so a bounce is identical every time. Detune (0 to 50 cents) and Spread set how far apart they sit.
Drop (0 to 60 st) starts the grind above the played note and falls back to it; Time (2 to 200 ms) is how long that takes. Deep and short is a beater; a few semitones is a bend into the note. Time is greyed out until Drop is up.
Drop is the grind’s half of a kick. Pair it with Knock in the SUB rail, which does the same to the sub: with only the sub falling, a short note still reads as a short note, because the grind’s harmonics hold the pitch still. With both, the whole sound falls at once.

Cutoff (-12 to +60 st): the interval above the played note; the readout beneath names it.
Resonance: the peak’s height, with the interval readout beside the value. At the top of the range the filter self-oscillates, in tune with the note.
Keytrack (0 to 200%): how much the cutoff follows pitch, around C2. 100% is a fixed interval above every note; 0% parks the filter where a C2 would have it, like an ordinary fixed filter; 200% opens the filter faster than pitch rises.
Drive (0 to 24 dB): level into the filter’s nonlinearity. More drive thickens and compresses the resonance without pulling it out of tune.
The pluck lives in the MOD row: the Env source’s Cutoff depth, in semitones (-60 to +60), is the distance the envelope moves the cutoff, so the same patch plucks by the same interval on every note.

Flavor: Clip, Tanh, Sine Fold, Linear Fold, or Cubic. Clip and Tanh saturate, the folds wrap the waveform back on itself for harsher spectra, and Cubic is the softest of the five: it adds a third harmonic and little else until it is driven hard.
Drive (0 to 36 dB), Bias (static asymmetry, adds even harmonics), and Mix (dry/wet).
The shaper runs on each unison voice separately, before they are summed, so detuned voices do not beat against each other inside the distortion. Because the filter is keytracked, the signal entering the shaper has the same shape in every register, so Drive feels the same on every note.

Cleanup: distortion regenerates energy at the fundamental, which would stack against the sub and make the low end depend on the shaper settings; this stage removes it. In Track mode a notch sits exactly on the sounding note, glide included, with a 12 dB/oct highpass just under it. Fixed mode is the highpass alone, with a Hz control (20 to 400), for when you want the grind to keep some of its own low end.
Chorus: two voices. Rate, Depth, and Mix. Stereo detunes the left and right sides in opposite directions, so the width comes from a pitch difference between the ears rather than the Widener’s time difference; off, the chorus thickens a mono grind without widening it. The two stack.
Widener (0 to 20 ms): a one-sided delay. Blend sets the delayed side’s place in the image: at 0 the grind is mono at full level, at 100 you get the full offset. If a mono check collapses your patch, pull Blend back before touching anything else.
Width order is fixed: unison, then chorus, then widener. None of it exists on the sub path.