endogen

Overview

ENDOGEN is a modular lowercase synthesis environment built around microsound, drift, feedback, and near-silence. It includes two generative oscillators and ten additional oscillators with direct frequency control, all of which can be individually tuned and operate within constrained pitch structures, featuring octave scale shifting and continuous frequency control.

Beyond this, ENDOGEN includes 14 complex modal synthesis modules modeled after vactrol-like dynamics and the frequency responses of tapes, cabinets, and amplifiers. These modules do not aim for faithful emulation, but for behavioral resemblance: slow energy transfer, asymmetric decay, spectral bending, and subtle saturation. The result is a synthesis layer with a distinctly organic, quasi-analog character, capable of producing textures that feel physically grounded, unstable, and alive—especially at very low amplitudes, where timbral nuance becomes the primary musical material.

Designed for Listening, Performance and Recordings

ENDOGEN is built for live performance and studio work. Its architecture supports extended listening sessions, real-time manipulation during performances, and direct recording of its output for use in compositions and sound design projects.

~150 continuous parameters
Small numeric changes affect long-term behavior rather than immediate results.

Physical System Approach

ENDOGEN is designed to behave like a physical system rather than emulate one. Its synthesis engine is particularly suited to modeling the kinds of signals you get from contact microphones, pickups, and field recordings focused on material resonance. Instead of replaying or filtering recordings, ENDOGEN works on excitation, friction, and energy transfer, allowing resonant bodies to emerge from noise, impulses, and continuous pressure.

Material Simulation

This makes it especially effective for simulating:

• contact mic interactions with wood, metal, glass, membranes, and cavities
• unstable resonances typical of piezo pickups and surface transducers
• low-level feedback and micro-vibrations found in close-range field recording

Liquid-like Behaviors

Beyond solid materials, ENDOGEN can also articulate liquid-like behaviors. By shaping turbulence, friction, and airflow-like excitation, the system can suggest:

• different liquid densities and viscosities
• variations in flow speed and friction
• the perceived size of openings or orifices through which air or liquid is forced
• bubbling, gurgling, splashing, and breathy pressure states that continuously morph rather than loop

Rather than discrete events, these behaviors exist on a continuum: friction becomes tone, noise becomes resonance, and resonance collapses back into instability. At very low amplitudes, this produces soundscapes that feel tactile, intimate, and physically present — closer to listening through a contact mic than to traditional synthesis.

A Max-SuperCollider Hybrid

ENDOGEN is controlled by a Max interface, but the main audio engine is in SuperCollider for very practical reasons:

• Accumulation of code and sonic research
  Over the past two years I've studied Supercollider in depth and accumulated many synthesis "pieces" that, when listened to, felt closer to the kind of behavior and timbre I was looking for: a very direct quality, sometimes even a bit "wrong" (in a good way), with a response that recalls certain LPG-type articulations and certain modular habits.
• Unification in a single system
  At a certain point I did a "collect & save" job: thousands of lines, one single sound patch, and then wiring between sections to create interactions where needed. Not everything is interdependent, but some areas are in subtle ways.
• Better user interface (Max) without giving up sound (SC)
  SuperCollider is extremely powerful, but its standard interfaces are visually harsh and not very ergonomic. Max instead is unbeatable as an interface and control tool. So: SC for the sound, Max to make the instrument usable.
• It's not ideology, it's project choice
  Max remains central and for me it's an irreplaceable way of thinking. Here SC was chosen because that engine, for this specific project, gives me exactly the kind of sonic "matter" I wanted to preserve.

Technology note (Max ↔ SuperCollider)
Communication between Max and SuperCollider happens via OSC. After installing a small set of dependencies, the launcher automatically starts a lightweight SuperCollider server in the background. The engine status can be monitored inside the Max interface, while a parallel macOS Terminal session—opened automatically by the launcher—provides a clearer, more fluid readout of system activity during use. No manual interaction with SuperCollider is required.

electroacoustic lowercase synthesis

endogen is a synthesis environment designed for electroacoustic lowercase music and microsound: very low levels, subtle dynamics, and details that emerge slowly. If you’re drawn to Steve Roden’s micro-event aesthetics or reduced listening practices within minimal electroacoustic traditions, you’re in the right territory.

endogen is not a preset machine and it’s not meant to deliver instant results. It is an instrument for slow listening: patience changes everything.

Listening first (recommended)

Before opening endogen, it helps to listen to a few works that define this territory. This isn’t about references for their own sake — it aligns expectations and makes the instrument easier to understand.

• Forms of Paper — Steve Roden
• Un Peu de Neige Salie — Bernard Günter
• Herbario — Federico Durand (12k) — and other releases on 12k
• Selected works by Thomas Köner (especially relevant for drone + sub/infra approaches)

A Reference:

In terms of lineage, ENDOGEN resonates with the ultra-quiet, material-focused sensibility of Steve Roden especially the way Forms of Paper treats fragile sound as a kind of closeup tactile field, where micro-events, air, and surface become the actual “composition.” That same attention to near-silence and minute behavior is central here: sound is not pushed forward, but allowed to emerge as a granular, intimate evidence of process.

A related point of reference is Bernard Günter’s Un Peu de Neige Salie, where extreme reduction, distance, and barely-there sound events create a fragile perceptual field. In that work, silence, residue, and spectral imbalance become compositional tools rather than absences. ENDOGEN shares this attitude: sound is treated as a minimal trace, shaped by slow internal processes and subtle instability, inviting a mode of listening focused on texture, threshold, and the physical presence of sound at its most restrained.

This kind of sonic attention — where small objects, quiet gestures, and fragile moments become the focus — is central to endogen's design philosophy.

Video

Watch demonstrations and sound examples from endogen on the official YouTube playlist.

Very low levels: what it means (and what you can do)

endogen runs very quietly by default. This is not a limitation: it’s a design choice that preserves micro-detail and keeps stable headroom when multiple modules are active.

Why start so low

• Preserve micro-transients and fragile detail
• Maintain headroom with many layers active
• Avoid spectral collapse when the bandwidth opens up
• Let infra-low, mid textures, and high micro-detail coexist without masking

Raising level is part of the process

endogen is meant to be turned up. Depending on your destination, you can raise monitoring level, use external soft compression, and do downstream gain staging. As you push level, the bandwidth becomes more explicit: infra-low movement turns physical, and high micro-detail reads as air and grain.

Exercise: Permafrost “shockwave” test

A practical way to understand endogen’s calibration is to test it with a simple, stable source.

1. Open Permafrost.
2. Select the first slider DRONE F.
3. Keep the system stable (few modules, no chaos).
4. Slowly raise your listening level.

That simple sine becomes a heavy, “teutonic” component of the environment. As you push it, you can perceive a kind of slow pressure wave — not because it is aggressive, but because endogen is tuned to hold energy over long timescales without breaking.

Context (why endogen exists)

We live in an era of extremely advanced VST instruments with massive technical possibilities and virtually unlimited palettes. That’s not a problem by itself — but it raises a basic question:

In earlier phases of electronic music, limited technology often made aesthetics more direct and easier to articulate. Today, it’s easier to get lost in possibility instead of making a clear choice. endogen responds to that condition: it doesn’t add more options — it narrows the field and commits to a specific territory.

What kind of technology endogen uses

endogen uses advanced technology, but not in a “spectacular” or futuristic sense. It is:

• behavior-oriented, not function-oriented
• systemic, not performance-first
• built for listening, not for demonstration

Under the hood there are complex structures, but the goal isn’t to show complexity. The goal is slow interaction: small changes in density, balance, and time create real differences that only become obvious through attention.

Expectations (important)

endogen is not a “wow in 5 seconds” instrument. It rewards slow listening, careful balance, and small changes over time.

What it is not (and why this helps)

• No presets. endogen behaves more like a bench instrument / tape setup than a recallable synth.
• No parameter-by-parameter manual. endogen works best when you explore modules as behaviors.
• Not fully cybernetic. Some areas are interdependent, others are intentionally more independent.

How to use it (the rule that avoids 90% of questions)

Open one module. Set density to minimum. Listen. Raise density slowly.

Many modules have a Density / Rate / Activity control — treat it as your macro lens.

• All modules OFF (or at zero output if no on/off)
• Enable one module
• Start with very low density
• Raise density gradually (touch nothing else)
• Only then adjust tone / filtering / dynamics
• Add a second module and repeat

Working Slowly with Detail

To generate lowercase soundscapes, work on modules very slowly, keeping the volume of elements low. While you can obviously reach high volumes, creating lowercase textures requires focusing on the small and the detail. All endogen modules are designed for this: from complex noise generators to organic modal synthesis to vactrol LPG synthesis.

Try to model small objects falling, hitting, and resonating on different surfaces. Think of marble impacts, soft wooden taps, glass fragments shifting — micro-events that reveal themselves through careful attention and patient listening.

Patience: what to expect in the first minutes

• Many changes are micro and happen over long timescales
• After 10 seconds it may feel like “nothing is happening”
• In practice, endogen often works through slow emergences

In the lowercase context, this is intentional.

Randomization and organic LFO

Many parameters offer two approaches: randomization (discrete “dice” changes) or an organic LFO (slow, non-periodic continuous motion).

The organic LFO is built in gen and behaves closer to irregular drift than classic sine LFOs. It is conceptually related to the Ghost LFO used in my other tool, Envion (brief mention only).

Resonant Objects: LFO, randomization, TRIGGER

Resonant Objects is built around a complex triple oscillator feeding a vactrol-modeled Low Pass Gate. The result can move from dry, glassy impacts (marble-like) to soft percussive LPG behavior (Buchla-style “bongo” articulation).

Randomization vs LFO (in this module)

Here, randomization is mainly for fast exploration: generating dozens of configurations (including fine-tuning variants) to map the range of resonances the module can create. The organic LFO is for long, coherent evolution once you’ve found a region you like.

TRIGGER (daily sound design)

With single triggers or very low density, you can generate isolated hits. This supports a quick collect & save workflow: generate → record → reuse anywhere outside endogen.

Complex VCO & Mechanics

ToyVCO — FM Oscillator

ToyVCO is a three-operator FM oscillator with ladder filter control, resonance, and FM-of-FM routing. It's capable of generating everything from small, sweet, slightly lo-fi toy-like sounds, reminiscent of children's instruments, to striking basslines with a sharp, pronounced bite.

The Mechanics Section — Beyond Simple Tape Emulation

In the video below, you can also clearly hear the depth of the Mechanics section. This is not a simple tape emulation. I was extremely demanding here.

The system reproduces warm dual-stereo characteristics, alongside a capstan modeled on a solenoid that already introduces frequency inversion.

As speed increases, low frequencies grow exponentially, becoming powerful and almost oppressive. When slowing down, the low end gradually thins and depletes.

There is also dedicated control over capstan flutter and an equalizer with inverted phase. Combined with the deck section, the entire object modeling turns into an ancient machine, operating at a very high level of abstraction.

From Fragile to Rural-Industrial

This video shows clearly how endogen can move effortlessly from subtle, fragile soundscapes to a rural-industrial sonic environment, without breaking coherence.

Cybernetic areas

endogen is partly cybernetic: some regions self-influence through simple feedback relations. This becomes more evident when balancing higher-energy / noise modules alongside other layers. These behaviours open a space for a deeper exploration of cybernetic thinking, in particular the work of Roland Kayn and Jaap Vink, and of feedback understood not merely as an effect, but as a generative and structural principle shaping the system’s long-term behaviour.

• Air Pressure
• Cybernet
• Mstnagra
• Herbario

Synthesis Modules

ENDOGEN includes 30+ specialized synthesis modules, each designed to generate specific sonic behaviors. These modules are not presets but distinct synthesis engines, capable of producing textures ranging from microsound to large-scale emergent patterns.

Core Synthesis

Generative Oscillators

Modal & Resonant Synthesis

Cybernetic Systems

Tape & Mechanical Systems

Noise & Textural Generators

Spatial & Environmental

High-Frequency Layers

Processing & Utilities

Core architecture (overview)

Companding (dynamic balance)

A central part of endogen’s sound design is an extensive companding system (upward + downward compression). Instead of flattening the mix, it redistributes dynamic energy so ultra-quiet micro-events, mid textures, and deep low-frequency structures can coexist without masking each other. The goal is a dense but transparent field that remains stable over long durations.

Monitoring

endogen is designed for full-range monitoring. Use good headphones or proper speakers/sub to catch the infra-low movement and the high micro-detail safely.

Development Journey

During the development of this system, I had to find a way to make the pure coding environment communicate, and the only robust and viable path turned out to be Open Sound Control from Max/MSP. So I designed a fairly substantial interface to articulate the timbre of the coding side with very little effort, and above all oriented toward performance and recording.

This approach bypasses the classic method of real-time evaluation and compilation in SuperCollider. Of course, this required a significant effort, mainly to understand not only the automation logics (gen~ DSP and pure Max), but also levels and scales, in order to make the whole system sound coherent in terms of frequency response.

I can finally say that I managed to achieve my goal. I have been experimenting with modular synthesis for years, and I am a huge enthusiast of vactrol circuits and field recording. For this reason, I always tried to benchmark at home, carefully evaluating the timbral response of the low pass gates I own.

It was an exhausting process for me, living secluded in my own small world.

About the Author

endogen is developed by Emiliano Pennisi, a contemporary sound artist and composer. Learn more about his work and projects.