Vector Space is not a pitch source by itself. It’s a CV relationship generator: you feed it three modulation or pitch-related voltages, and it gives you 17 related outputs. That makes it extremely useful for building melodies, harmonies, counterpoint, voice-leading, and evolving pitch structures.
Vector Space takes three inputs: i, j, k.
It then derives 17 outputs in three families:
It can process CV or audio, but for melody-making the key idea is:
Send in three changing voltages, then use the different outputs as related melodic control voltages.
Because all 17 outputs are derived from the same three sources, they tend to sound coherent together, even when they are different enough to create separate melodic lines.
Each input has a ++ / +- switch:
This matters a lot for melodic work.
If you’re patching into a quantizer, these switches are a great way to push a line into a different register or change its contour.
The Cube outputs are the most straightforward combinations of the three inputs, with in-phase and out-of-phase mixtures.
These outputs are easier to understand and tend to preserve the identity of the incoming CVs. If your three inputs are: - a sequence - a slow transposition source - a rhythmic stepped random
then each Cube output becomes a different weighted/sign-flipped combination of those three behaviors.
Run several Cube outputs into separate quantizers or into a multi-channel quantizer, then send them to multiple VCOs. This gives you: - one melody - a countermelody - a bassline - a harmony voice
all evolving from the same underlying motion.
The manual specifically points out a special property of the Plane outputs:
at any one moment, three of the outputs will have similar voltages, but which three changes as the input phases change
That is extremely musically useful.
Because these outputs use rectified versions of two inputs plus a phased version of the third, they are more skewed positive and can create very interesting pitch behavior after quantization.
When quantized, the “three outputs tracking together” behavior can produce: - temporary unisons - close intervals - shifting triad-like movement - melodic braiding between voices
This is especially effective when each Plane output goes to its own oscillator voice.
Rectification can effectively double modulation frequency, so if one of your inputs is an LFO or oscillating stepped source, Plane outputs can generate more active note changes than the original source.
These three are less like “normal mixed pitch CVs” and more like meta-melodic controls.
A sum of all rectified inputs, representing how far the point is from the center.
Use it for: - a master transposition lane - melody intensity - opening into higher register as motion increases - quantized “energy melody”
Because it’s based on rectified inputs, it tends to stay positive and can create a strong upward-biased melodic contour.
Negative version of Sphere.
Use it for: - contrary-motion bassline - opposite-register melodic reflection - downward counterline to a Sphere melody
This is especially useful if Sphere controls a high voice and NegSphere controls a bass voice.
Represents distance to the closest edge of the sphere; highest at the center, lower toward the edges.
Use it for: - central “home note” behavior - tonic-like stabilizing melody - drone pitch selection - a middle voice that is most active when the modulation space is balanced
This output is especially good for creating a sense of melodic center while the Cube and Plane outputs create more unstable motion around it.
This is the most obvious melodic use from the manual.
You get several related melodic voices that: - often move together - then split apart - then regroup in new combinations
This can create rich ambient, generative, or minimalist melodic systems.
Use Vector Space as the center of a multi-voice patch.
A full melodic ecosystem from one sequencer plus two modulation sources.
The sequencer provides recognizable musical structure, while the other two sources constantly reshape interval relationships.
This works well if you want music that sounds composed rather than random.
Use several Cube outputs into quantizers and then VCOs.
The different outputs will behave like alternate harmonizations of the same motion. Since the outputs are mathematically related, your voices tend to move with internal logic rather than feeling disconnected.
This is great for: - Berlin-school lines - modal harmony - evolving ostinati - contrapuntal polysynth-style patches
The manual discusses joystick use for spatialization, but it also works beautifully for melody.
Take several Cube/Plane outputs through quantizers to oscillators.
Your hand movement “navigates” a field of related notes and harmonies. This gives a very performable way to: - sweep through chord voicings - discover melodies by hand - improvise harmonic changes - control several voices at once
UnSphere is especially useful here as a center-focused voice.
Instead of using all outputs directly as pitch, use some of them to control the melody system around a main sequence.
The melody stays anchored by a clear sequence, but the harmony and accompaniment breathe around it.
This is often more musical than using all 17 outputs as raw pitch.
Best for: - generative melodies - multi-voice pitch generation - counterpoint
Examples: - sequencer + random + sample & hold - three differently clocked sequencers - one sequence copied into two attenuated/offset variants plus a third random source
Best for: - melodic patches that feel structured but alive
This is probably the sweet spot for many musical uses.
Best for: - looping melodic systems - phase-based canon patterns - pseudo-polyrhythmic melodies
Since Plane outputs can frequency-double through rectification, this can make surprisingly complex note movement from simple LFOs.
Best for: - timbral/audio use more than conventional melody
But if heavily attenuated and quantized, audio-rate interactions can create unstable stepped melodic artifacts. More experimental than tonal.
For pitch work, a very useful chain is:
CV sources → Vector Space → attenuation/offset if needed → quantizer → VCO 1V/oct
Why: - Vector Space can output large voltages - Some outputs may benefit from scaling before quantization - Quantizers turn its complex CV relationships into musically usable note sets
If multiple outputs are quantized to the same scale, the voices stay musically related even when the raw CVs diverge.
Good scales: - minor pentatonic for forgiving generative patches - dorian/aeolian for melodic ambient work - chord quantization for strong harmonic identity
Even if two outputs are close in voltage, placing oscillators in different octaves makes the result feel intentional and layered.
If some outputs are too jagged: - keep one voice fully stepped - slew another slightly - leave a third un-slewed but quantized
This creates more expressive melodic contrast.
Since the module has several output families, a very musical approach is to assign them roles:
Use for: - lead - alto - tenor lines
Use for: - clustered voices - inner parts - accompaniment arps
Use for: - bassline - transposition - tonic center - phrase intensity
This division gives the patch a natural musical hierarchy.
Set: - sequencer likely unipolar - random source depends on source, often unipolar - LFO usually bipolar
This is a very efficient way to get a complete melodic composition from a small number of CV sources.
You can “steer” through harmonically related note regions by hand. This is excellent for: - live improvisation - soundtrack work - ambient chordal movement - controlled generative melody
Vector Space works especially well with:
If your goal is melody, the most important companion is a good quantizer.
In a melodic patch, Vector Space is best thought of as one of these:
It is less about writing one exact melody and more about creating a field of related melodic possibilities.
WORNG Vector Space is excellent for creating melodic components when used with quantizers and multiple voices. Its strength is not “a single tune,” but families of interrelated pitch lines:
If you feed it: - one structured source, - one semi-random source, - and one slow modulator,
you can get a whole melodic arrangement out of a single patch.