Free 32-Partial Additive Synthesizer
Overtonium is a free and open-source additive synthesizer by Benjamin Dehli / Dehli Musikk that takes a refreshingly direct approach to additive synthesis: instead of hiding harmonics behind a spectrum editor, it lays 32 individual partials out like the channels of a mixing console.
Every channel is a sine oscillator tied to a harmonic of the note you play, and every partial gets its own tuning, envelope, pitch modulation, amplitude modulation, drift, velocity response, aftertouch, pan and level.
There is also a dedicated noise channel.
The result is less like programming a conventional synthesizer and more like mixing the internal structure of a sound.
Overtonium is completely free, released under the AGPLv3 open-source license, and available for macOS, Windows and Linux.
Additive Synthesis as a 32-Channel Mixer
Additive synthesis works by constructing complex sounds from simpler sine waves.
A single sine oscillator produces an almost perfectly pure tone.
Add more sine waves at different harmonic frequencies and amplitudes, and increasingly complex timbres begin to emerge.
Overtonium gives you direct access to 32 of those partials.
Instead of drawing their amplitudes into a spectral graph, each harmonic receives its own mixer channel.
That means shaping a sound becomes something remarkably intuitive:
you mix its harmonics.
Want more fundamental?
Raise channel 1.
Want more octave?
Adjust partial 2.
Need more fifth?
Work with partial 3.
Want something considerably stranger?
Start changing the tuning, envelopes and modulation of individual harmonics independently.
This is where Overtonium moves far beyond a simple additive oscillator bank.
32 Partials + Noise
The main sound engine consists of:
- 32 harmonic partials
- 1 independent noise channel
Each partial receives its own complete signal-shaping channel.
The interface therefore resembles a huge mixing desk rather than a conventional synthesizer panel.
And there are a lot of controls.
Each of the 32 harmonic strips contains 23 parameters, with additional controls on the noise channel and global processing section.
In total, 784 parameters are exposed to the host for automation.
Every Partial Can Behave Differently
Each harmonic channel provides control over areas including:
- Tuning
- Phase
- Pitch modulation
- Drift
- Envelope
- Key-off behavior
- Amplitude modulation
- Velocity
- Aftertouch
- Pan
- Mute
- Solo
- Level
This allows different portions of the harmonic spectrum to behave almost like independent instruments.
For example, the fundamental can attack immediately while higher harmonics gradually fade in.
Odd harmonics can tremolo while even harmonics remain static.
Upper partials can react strongly to velocity while the fundamental remains stable.
One harmonic can wander randomly in pitch while another remains perfectly tuned.
This makes Overtonium capable of producing spectra that evolve dynamically rather than remaining fixed.
TUNE — From a Timbre to a Chord
One of Overtonium’s most distinctive features is TUNE.
Normally, additive synthesis places partials at exact integer multiples of the fundamental.
Overtonium allows every partial to move continuously between:
Equal temperament
and
Just intonation.
At the just-intonation end, the partials sit at exact whole-number ratios with the fundamental.
The stack therefore behaves like the harmonic spectrum of one complex sound.
Move toward equal temperament and each partial approaches the nearest chromatic semitone.
Suddenly, what was perceived as one timbre begins to separate into something closer to a chord.
The fascinating part is that the synthesis structure has not changed.
Only the tuning relationship between the partials has.
Six Keyboard Temperaments
The tuning of the harmonic spectrum and the tuning of the keyboard are treated separately.
Overtonium includes six keyboard temperaments:
- Equal
- Just
- Pythagorean
- Quarter-comma meantone
- Werckmeister III
- Young
The temperament can be built around any of the twelve roots.
Reference pitch can also be adjusted from:
415 Hz to 466 Hz
This makes Overtonium particularly interesting for musicians interested in historical tuning systems, just intonation and microtonal experimentation.
STRETCH — Break the Harmonic Series
Real acoustic instruments do not always produce perfectly harmonic overtones.
Strings with physical stiffness, for example, produce progressively sharper upper partials.
Overtonium’s STRETCH control recreates this type of inharmonicity.
At moderate settings, the effect can resemble the stretched harmonic structure associated with a piano.
Push it further and the partials increasingly stop agreeing on a common fundamental.
The sound begins moving away from a conventional pitched instrument toward:
- Bells
- Metallic tones
- Inharmonic textures
- Experimental percussion
Negative Stretch values are also possible, even though that behavior does not correspond to a normal physical string.
That is deliberate.
Overtonium is perfectly willing to move beyond physical reality when the result is musically useful.
TRACK — Make the Spectrum Change Across the Keyboard
A static additive spectrum can sound unnatural because real instruments rarely maintain exactly the same harmonic balance at every pitch.
Overtonium addresses this with TRACK.
TRACK progressively rolls off upper partials as you play higher notes.
The fundamental remains stable while the upper harmonic structure becomes thinner.
This means a bass note can retain a rich overtone structure while the same patch played several octaves higher becomes naturally softer and less harmonically dense.
DRIFT — Every Partial Can Wander
Each harmonic has its own DRIFT control.
DRIFT introduces a smooth random pitch walk of up to:
25 cents
Importantly, every partial uses its own random stream.
The result is not conventional global vibrato.
The individual components of the sound slowly disagree with one another.
A sustained chord can therefore continue evolving even when no conventional modulation is active.
This feature traces its heritage to Benjamin Dehli’s earlier cassette-based instruments, where natural tape instability created constantly changing pitch relationships.
Per-Partial Pitch Modulation
Every partial also receives its own dedicated pitch modulator.
Available rates extend from:
0.01 Hz to 30 Hz
with modulation depths reaching:
1,200 cents
— a complete octave.
Available modulation shapes include:
- Sine
- Triangle
- Sawtooth
- Reverse Sawtooth
- Square
- Sample & Hold
- Random
Pitch modulation can therefore range from subtle vibrato to radical harmonic reorganization.
Because every partial can use different settings, extremely complex movement is possible.
Per-Partial Amplitude Modulation
Each harmonic also receives its own amplitude modulator.
Again, different shapes and rates can be assigned independently.
This allows one harmonic to pulse slowly while another gates rhythmically and another remains completely static.
Instead of modulating the volume of the whole synthesizer, you are modulating the internal spectrum itself.
That distinction is fundamental to what makes Overtonium interesting.
Detailed Envelopes for Every Harmonic
Every partial has its own envelope.
Controls include:
- Strike
- Delay
- Attack
- Decay
- Sustain
Attack can be as short as 0.2 ms, while decay can extend to 20 seconds.
Delay can reach 5 seconds.
By staggering delays across the harmonic series, a sound can gradually construct itself over time.
The fundamental might begin immediately.
The third partial arrives later.
Then the fifth.
Then progressively higher harmonics.
Instead of simply fading in, the timbre itself unfolds.
Independent Key-Off Envelopes
Overtonium also provides a separate envelope for what happens when you release a key.
Controls include:
- Swell
- Level
- Release
This allows unusual release behavior that conventional ADSR envelopes cannot easily reproduce.
A harmonic can disappear normally, bloom after key release or even briefly become louder.
This can recreate behaviors associated with mechanical instruments while also opening up much more experimental possibilities.
Velocity Changes Timbre, Not Just Volume
Velocity is independently adjustable for every partial.
This means playing harder does not simply make the whole patch louder.
You can configure higher harmonics to respond more strongly than the fundamental.
A soft note might therefore sound dark and rounded.
Strike the same note harder and upper harmonics emerge, making the timbre brighter and more aggressive.
Negative velocity response is also available.
That allows two groups of harmonics to crossfade against one another according to playing strength.
Per-Partial Aftertouch
Aftertouch is also available independently for each harmonic.
Positive values bring a partial forward as pressure increases.
Negative values can make it disappear.
This becomes particularly powerful with expressive MIDI controllers.
MPE Support
Overtonium supports MPE.
With MPE enabled, pitch bend and pressure can operate independently for individual notes.
This means pressing one note harder within a chord can alter only that note’s harmonic spectrum.
The MPE Slide dimension can also control:
- Brightness
- Tuning
For example, one finger can move a note between equal temperament and just intonation while the remaining notes of the chord remain unchanged.
The Noise Channel
Alongside the 32 harmonic oscillators is a dedicated noise channel.
It receives many of the same controls as the partials, including:
- Envelope
- Tremolo
- Velocity
- Aftertouch
- Pan
- Mute
- Solo
- Level
Instead of harmonic tuning, it provides a COLOUR control.
This moves continuously from dark low-frequency rumble through flat noise to bright hiss.
Each voice also receives its own independent noise stream.
LINK — Control 32 Channels Without Losing Your Mind
Thirty-two harmonics with dozens of parameters each could easily become unmanageable.
Overtonium solves this with an intelligent LINK system.
LINK allows the same parameter across multiple partials to be adjusted simultaneously.
But unlike a conventional parameter link, it does not simply force every control to the same value.
Changes are applied relatively, preserving the spectrum you have already created.
Several scopes are available:
- All partials
- Same interval
- Odd harmonics
- Even harmonics
Different response curves can also be used:
- Uniform
- Taper from the selected harmonic
- Spread / Gather
This makes it possible to reshape a complete harmonic structure with a single gesture.
Oscillator Character
Although every partial begins conceptually as a sine oscillator, Overtonium includes several oscillator characters:
- Pure
- Bulb
- Rail
- Diode
- Valve
- Op-amp
These are not simply conventional distortion effects applied after the synthesizer.
They alter the behavior of the individual oscillator bank.
The different characters introduce behaviors inspired by imperfect electronic oscillator circuits, including harmonic additions, nonlinearities and small fixed deviations between partials.
That gives the otherwise mathematically clean additive engine a more physical personality.
Global Processing
The instrument also includes global processing after the 32-channel mixer.
The current architecture includes tools such as:
- Wobble
- Tape-style echo
- Reverb
- Master level
- Safety clipping
These allow the highly detailed additive source to be shaped into a more complete finished instrument without immediately requiring external effects.
Polyphony
Overtonium supports:
- Legato
- 1 voice
- 2 voices
- 4 voices
- 6 voices
- 8 voices
- 12 voices
- 16 voices
The default is 8 voices.
Because each voice can contain 32 oscillators, eight-note polyphony means the engine may be running 256 harmonic oscillators simultaneously.
At sixteen voices, that increases to 512 oscillators.
Surprisingly Reasonable CPU Usage
Despite the architecture, the developer reports relatively modest CPU requirements on the reference system.
At 48 kHz with all modulation, drift, metering and noise active:
1 voice / 32 oscillators: approximately 1% of one CPU core
8 voices / 256 oscillators: approximately 7%
16 voices / 512 oscillators: approximately 14–15%
The global tape echo and reverb add less than approximately 1% on that test system.
Naturally, actual performance will depend on your computer.
Reduce the Render Rate for Lo-Fi Sounds and Lower CPU
Overtonium can also reduce the internal render rate.
This is not simply a filter placed over the output.
The complete voice engine actually operates at the lower rate.
At 8 kHz, the developer reports that the voice pool costs approximately one fifth as much CPU.
And because the reduced rate is audible, the control doubles as a creative lo-fi sound-design tool.
Presets
Overtonium includes a collection of factory patches demonstrating different sides of the synthesis engine.
The presets are particularly useful because an instrument with 32 individually programmable harmonics can initially look intimidating.
They provide starting points for understanding how tuning, harmonic balance, envelopes and modulation interact.
Full Host Automation
A remarkable 784 parameters are exposed to the host.
This includes:
- 30 global parameters
- 23 parameters × 32 partials
- 18 noise-channel parameters
Even global settings such as temperament, polyphony and pitch-bend range can be automated.
That creates some genuinely unusual possibilities.
A composition could, for example, transition gradually from one tuning temperament into another.
Free and Open Source
Overtonium is not merely free of charge.
It is free software.
The complete source code is available publicly under the:
GNU Affero General Public License v3 — AGPLv3
The project is built with JUCE 9.
There is no commercial unlock, reduced free edition or paid upgrade required to access the synthesis engine.
Version 1.9.0
At the time of writing, the current release is:
Overtonium 1.9.0
released on:
September 23, 2026
Installers and portable ZIP packages are available depending on platform.
macOS
macOS receives:
- VST3
- Audio Unit
- Standalone
The build is Universal and supports:
- Apple Silicon
- Intel
Minimum versions are:
macOS 10.13+ on Intel
macOS 11+ on Apple Silicon
The macOS installer is signed and notarized.
Windows
Windows receives:
- VST3
- Standalone
Both an installer and ZIP package are available.
The Windows installer is currently unsigned, so Windows SmartScreen may identify the publisher as unknown.
Linux
Linux receives:
- VST3
- LV2
- Standalone
There is no installer.
Instead, the download provides the required plugin files directly.
An Important Note About Windows and Linux Testing
The developer is unusually transparent about platform testing.
Automated builds and tests run across all three supported platforms, and VST3 builds are tested with pluginval.
However, the developer states that, as of the current documentation, the Windows and Linux builds have not yet been tested inside an actual DAW by the developer.
macOS is the platform on which the VST3 and Audio Unit versions have additionally been tested in real hosts.
That does not mean the Windows or Linux versions will not work, but it is useful information for an early release.
What Can Overtonium Make?
The architecture is capable of much more than textbook additive synthesis.
It can be particularly interesting for:
Evolving Pads
Stagger harmonic envelopes and use independent modulation to make the spectrum continuously evolve.
Bells and Metallic Sounds
Use STRETCH to break the strict harmonic relationship between partials.
Organs
Treat the partial faders like an enormously expanded drawbar system.
Experimental Drones
Combine Drift, modulation and long envelopes across the harmonic spectrum.
Microtonal Music
Experiment with just intonation and historical temperaments independently from the partial tuning system.
MPE Performance
Use pressure and slide to reshape the harmonic spectrum independently for each note.
Digital and Lo-Fi Textures
Reduce the internal render rate and combine it with oscillator character and modulation.
Unusual Chords
Move the harmonic series toward equal temperament and let what was originally one timbre begin separating into chromatic intervals.
Physical-Like Sounds
Combine Stretch, Track, velocity-sensitive partials and oscillator character to create behavior reminiscent of acoustic and electromechanical instruments.
Strengths
- Completely free
- Open source
- AGPLv3
- 32-partial additive synthesis
- Dedicated noise channel
- Mixer-style workflow
- Independent level for every partial
- Independent tuning
- Equal-to-just TUNE control
- Inharmonic STRETCH
- Spectral keyboard tracking
- Per-partial phase
- Per-partial pitch modulation
- Per-partial random Drift
- Per-partial amplitude modulation
- Independent envelopes
- Independent key-off envelopes
- Per-partial velocity
- Per-partial aftertouch
- Independent pan
- Mute and Solo
- Intelligent LINK system
- Odd/even harmonic grouping
- MPE
- Polyphonic aftertouch
- Six keyboard temperaments
- Twelve temperament roots
- 415–466 Hz reference tuning
- Six oscillator characters
- Up to 16 voices
- Adjustable internal render rate
- Global effects
- 784 automatable parameters
- Factory presets
- VST3
- Audio Unit
- LV2
- Standalone
- Apple Silicon
- Intel Mac
- Windows
- Linux
- Signed and notarized macOS installer
Potential Limitations
Overtonium is not designed for producers who want a conventional subtractive synthesizer with three oscillators, a filter and a simple ADSR.
There is no traditional workflow where you choose a saw wave, close a low-pass filter and immediately obtain a familiar analog patch.
The 32-channel architecture is intentionally detailed.
That can make the instrument initially intimidating.
There is also an unavoidable relationship between polyphony and CPU consumption: every additional voice can require another 32 oscillators and their associated modulation.
The developer has optimized the engine considerably, but extreme polyphony and complex modulation will naturally require more processing power than a simple synthesizer.
Windows users should be aware that the installer is unsigned.
And, as of version 1.9.0, the developer states that Windows and Linux builds have passed automated testing but have not yet been personally tested inside real DAWs.
These limitations aside, the architecture is precisely what makes Overtonium different.
Final Verdict
Overtonium is one of those free synthesizers that exists because somebody had a genuinely different idea about how synthesis should be approached.
Additive synthesis itself is nothing new.
What is unusual is the decision to expose the harmonic spectrum as a 32-channel mixing console and then give every one of those channels the tools normally reserved for an entire synthesizer voice.
You are not merely deciding how loud the 7th harmonic should be.
You can decide when it appears.
How it responds to velocity.
Where it sits in stereo.
Whether it drifts.
How it vibrates.
How it behaves when you release the key.
And whether pressure brings it forward or makes it disappear.
Then Overtonium adds another layer with its tuning system.
Moving between equal temperament and exact harmonic ratios can transform the same oscillator stack from something perceived as one complex timbre into something approaching a chord.
STRETCH can then pull the harmonic series apart completely.
Add historical temperaments, MPE, per-partial modulation, oscillator character, global effects and 784 host-automatable parameters, and this becomes a remarkably deep experimental instrument.
It may take considerably longer to understand than the average free synth.
But that is also the point.
Overtonium does not hide additive synthesis from you. It gives you the entire harmonic structure and asks you to mix it.
For sound designers, experimental electronic musicians and anyone interested in harmonics, alternative tuning or evolving timbres, Overtonium is absolutely worth exploring.
Sources :
benjamindehli.github.io: https://benjamindehli.github.io/overtonium/
https://www.audiopluginsforfree.com/wp-content/uploads/2026/09/overtonium-v190.zip
