Exploratory Sound Creation and the Trap of Modular Synthesis

The modular synthesizer, invented in the mid-20th century and experiencing a resurgence in the 21st, is widely regarded as one of the most powerful and flexible tools for electronic sound creation. Its defining promise is openness: arbitrary structures built from simple parts, assembled by the user rather than imposed by the instrument.

Yet despite this promise, modular synthesis remains trapped in a remarkably narrow conceptual frame.

There is a well-known taxonomy of modules with predictable behaviors and controls, such as oscillators, amplifiers, filters, ring modulators, and mixers. Some things to notice about these modules:

• they perform a single, simple behavior under normal operating conditions
• they touch only one or perhaps 2 sonic parameters (loudness, pitch, timbre…)
• they produce steady, time-invariant output when fed steady, time-invariant input
• they are not stateful under normal conditions
• their parameters are largely orthogonal
• they enforce distinctions between:
  • audio and control signals
  • continuous and event-based signals

These assumptions are rarely questioned. They are treated as neutral facts of synthesis rather than historical design choices.

Patching

Patching is the name given to programming a modular system by connecting various nodes to each other, traditionally using cables. Patching transforms a collection of unrelated modules into a signal topology.

At least two types of patching can be identified:

Compositional Patching

Compositional patching proceeds from a concept. The programmer begins with an idea of the structure that they want to build, even if only in outline. Modules are selected to execute an idea. Signal flow is hierarchical and directional. Roles are clear.

Modular synthesizers are optimized for this mode. Much canonical music created with modular synthesizers — whether associated with Wendy Carlos, Morton Subotnick, or contemporary artists — fits comfortably within this mode.

Exploratory Patching

Exploratory patching is fundamentally different. The programmer does not begin with a pre-conceived structural or sonic goal. Instead, they probe the system, establish relationships, and observe the behavior that emerges. The structure is discovered rather than specified.

This is not necessarily real-time performance, nor is it randomness for its own sake. It is a method of inquiry and discovery.

Modular synthesizers provide little structural support for this mode of exploration — an absence that artists like David Tudor implicitly recognized by working elsewhere.

The Narrow Funnel

Consider a typical attempt at exploratory patching:

• A sound source is required in order to get any sound at all. This is almost always an oscillator or noise source, producing a constant, high-energy output.

• This immediately implies the need for a VCA or filter in order to introduce dynamic volume.

• This immediately implies the need for one or more control signals to modulate that VCA or filter. These are almost always periodic (LFOs) or event-based (envelope generators triggered by clocks or sequencers).

Where is the exploration?

It occurs primarily in parameter space, not in structural space. Knobs are turned, ranges are scanned, but the underlying topology seems to form with little real choice from the user. The system strongly resists alternative organizational principles.

The modular synthesizer presents itself as open-ended, yet quietly funnels the user into the same few structural patterns over and over again.

At this point, a reader may protest that my example is too elementary; that interesting behavior would occur only if we use more complex topologies. But this objection already concedes the central problem: it assumes that a system must be composed before it can be explored.

The Deeper Issue

The deeper problem is not any specific module type, but a set of implicit assumptions:

• sound must precede structure
• control must be external to what it controls
• dynamics must be imposed, not intrinsic
• change must be triggered by events
• time must be segmented into cycles or pulses
• memory is rare, static, and typically does not degrade

These assumptions make exploratory patching difficult not because the user lacks imagination, but because the instrument itself continually collapses inquiry into execution.

What Would an Exploratory Instrument Require?

A system designed for exploratory patching would invert many of these assumptions.

Such a system would favor:

• relations over signals
• difference over sum (mixing)
• phase, alignment, or orientation over pitch
• continuous comparison over discrete events
• state and memory as first-class properties
• self-modifying behavior in addition to external modulation
• structures that behave meaningfully even in silence

In such a system, sound would not be the starting point. It would be an observation—one possible projection of an underlying process.

Exploration would occur at the level of topology, coupling, and feedback, not merely at the level of parameter adjustment.

A Necessary Clarification

It is important to acknowledge that instruments do exist which gesture toward some of the behaviors described here. Certain devices foreground interaction, instability, feedback, or non-orthogonal control, and can produce rich behavior with minimal or ambiguous structure.

However, these instruments tend to occupy the edges of the design space.

They are often:

• closed or semi-closed systems
• minimally or non-patchable
• deliberately unstable or chaotic
• opaque by design
• difficult to analyze, reproduce, or steer

Some excel at producing striking results, but not at supporting sustained inquiry.

What is being described here is not an extension of that lineage, but a missing center between several existing categories.

The Missing Center

On one side are traditional modular systems:

• highly general
• patch-programmable
• analyzable
• well-behaved
• but behaviorally inert unless carefully composed

On the other side are noise-oriented or anti-instrument devices:

• rich in emergent behavior
• non-orthogonal by default
• often intentionally obscure
• resistant to prediction or repetition

What is largely absent is an instrument class that is:

• behavior-forward rather than sound-forward
• patch-programmable rather than fixed
• structurally expressive at low complexity
• well-behaved without being trivial
• analyzable without being sterile
• capable of emergence without relying on chaos or obfuscation

This is not a call for unpredictability for its own sake. Nor is it an attempt to aestheticize malfunction.

It is an argument for instruments whose primary material is relation, difference, and coupling, and for which sound is only one possible manifestation.

Conclusion

Most modular systems are excellent at executing ideas and remarkably poor at generating them. They provide a powerful vocabulary for composition, but a very limited grammar for discovery.

The goal is not to escape structure, but to make structure itself the object of exploration. This is the space that SARC intends to occupy.