Adding VCV Rack 2 to the OAX Pergamon — Building a Principal 8′ Pipe from Scratch
Sometimes the best way to understand an instrument is not simply to play its sounds, but to build one of them yourself.
The WERSI Pergamon OAX1000 is already an extraordinarily capable instrument. It has three 76-note manuals, a pedalboard, drawbars, a VST3 host, sampling, FM, wavetable and analogue synthesis, as well as dedicated church-organ facilities. WERSI even describes the OAX system as having a three-oscillator integrated synthesizer.
So why would I want to add VCV Rack 2?
Because there is a considerable difference between selecting a synthesizer sound and building one.
VCV Rack turns the Pergamon into something rather different: a huge physical performance console connected to a virtual modular synthesizer in which I can decide exactly where the pitch comes from, what harmonics are present, how quickly the sound speaks, how stable its pitch is, how the transient behaves and what sort of acoustic space surrounds it.
And, rather appropriately, the first sound I want to build is not some enormous science-fiction pad.
I want to make a pipe organ.
More specifically, I want to create one of the foundations of the traditional organ:
The Principal 8′
Why Start with a Principal?
If you are trying to understand pipe-organ synthesis, a Principal is a particularly good place to start.
The Principal — often called a Diapason in British organ building — is neither trying to imitate a flute nor a string nor a trumpet.
It is essentially the characteristic organ tone.
A good Principal has a firm fundamental pitch, useful upper harmonics and enough brightness to give the note definition without becoming harsh. One Principal can sound wonderfully clear by itself, but Principals at 8′, 4′ and 2′ pitches also form the backbone of much traditional organ registration.
The designation 8′ tells us its pitch.
An 8′ stop sounds at normal written pitch.
Move to a 4′ stop and the pipes sound an octave higher.
A 16′ stop sounds an octave lower.
The name comes historically from the approximate length of the longest open pipe in the rank. It certainly does not mean that every pipe is eight feet long.
That immediately gives us our first design requirement:
Our VCV Rack oscillator must track the Pergamon keyboard at normal pitch.
But pitch alone will not make it sound remotely like an organ pipe.
The Pergamon Becomes the Controller
This is where the combination becomes particularly interesting.
The Pergamon is not merely a MIDI keyboard sitting in front of VCV Rack. It gives me three manuals, pedals, physical drawbars, expression pedals, programmable controls and a very organ-like environment in which to play the finished sound.
VCV Rack provides the modular laboratory behind it.
A simple conceptual signal path becomes:
Pergamon key -> MIDI -> MIDI-CV -> oscillator -> tone shaping -> VCA -> acoustic modelling -> audio output
VCV Rack's MIDI-CV module converts MIDI notes into the pitch and gate signals needed by a modular synthesizer, while Rack's Audio module provides the link between the virtual rack and an audio device. VCV Rack also supports polyphonic cables of up to 16 channels, meaning one modular signal chain can handle several simultaneously played notes rather than requiring a completely separate oscillator and filter for every key.
That is rather important for an organ.
Playing one note at a time would be somewhat limiting.
These official VCV Rack images also illustrate one of the things I particularly like about the software: unlike a conventional synthesizer where much of the signal path is hidden, here you can actually see where the sound is going.
Step One — Turning a Key into Pitch
The first module in my virtual organ is MIDI-CV.
The Pergamon supplies the MIDI note.
VCV Rack converts that into:
- pitch information;
- a gate telling the synthesizer that the key is being held;
- and, if wanted, velocity and other control information.
For an authentic pipe-organ Principal, however, I would deliberately not use key velocity to determine volume.
That is an important difference between a piano and a traditional mechanical or electric-action pipe organ.
Hit a piano key harder and the hammer strikes the string harder.
Press a conventional organ key harder and the pipe does not suddenly receive a violent gust of extra air.
The key essentially says:
Pipe on.
Then:
Pipe off.
So for this experiment I want the Pergamon to feel like an organ rather than a velocity-sensitive synthesizer keyboard.
That apparently small choice immediately makes the resulting instrument more convincing.
Step Two — The Oscillator: What Does a Principal Actually Look Like?
The obvious beginner's temptation is to choose a sine wave.
It produces a beautifully pure note.
Unfortunately, it also produces something rather unlike a Principal.
A sine wave contains only the fundamental frequency.
A real Principal pipe contains a series of harmonics. The exact balance changes according to pipe scale, construction, voicing, wind pressure and position within the rank, but broadly we want a sound with a strong foundation and gradually diminishing upper harmonics.
A sawtooth wave goes to the opposite extreme.
It contains a very rich harmonic spectrum and is usually far too buzzy when used untreated.
So my starting point would be to use the standard VCV VCO and experiment with blending outputs.
For example:
Triangle waveform: approximately 60-75%
Sawtooth waveform: approximately 25-40%
Those numbers are not sacred.
They are a starting point.
The VCV Free collection includes polyphonic VCO, VCF, VCA and ADSR modules, so the basic experiment can be constructed with a remarkably small set of modules.
The triangle provides the stronger, smoother body.
The saw introduces the upper harmonics that give the pipe some authority and presence.
Already we have moved from:
beep
towards:
organ tone.
But we are certainly not finished.
Step Three — Filtering the Harmonics
The oscillator produces our raw material.
Now we need to voice it.
This is where the VCF — Voltage Controlled Filter — becomes useful.
I would start with a low-pass filter.
The purpose is not to remove all the high frequencies. A Principal needs harmonic energy.
Instead, I want to tame the rather synthetic edge of the sawtooth component.
A useful first experiment would be:
VCO triangle + saw -> mixer -> low-pass VCF
Start with the filter fairly open.
Play middle C.
Gradually close it.
At first the sound will probably be far too bright.
Then it becomes increasingly rounded.
Close it too far and it turns into something approaching a flute.
Somewhere between those extremes is the character we are looking for.
This is one of those wonderful occasions when synthesis stops being a collection of knobs and begins becoming ear training.
You are asking:
What actually makes something sound like a Principal?
The Filter Should Probably Follow the Keyboard
There is another problem.
A fixed filter setting that sounds excellent around middle C may not work across the whole manual.
Low notes may become excessively bright.
High notes may become too dull.
Real organ pipes are not identical objects scaled mechanically from one end of the keyboard to the other. An organ builder voices individual pipes and can change scaling and construction across a rank.
Our simple electronic approximation needs some help.
One solution is filter keyboard tracking.
Part of the pitch CV can also influence the filter frequency.
As I move upwards through the keyboard, the cutoff rises slightly.
That allows upper notes to retain their harmonic character.
This is where modular synthesis becomes fascinating because suddenly we are no longer merely imitating the sound.
We are beginning to reproduce some of the behaviour of the instrument.
Step Four — An Organ Pipe Does Have an Envelope
It is sometimes said that an organ has no envelope because the sound stays on for as long as the key is held.
That is not quite true.
The steady-state part may remain approximately constant, but a real pipe takes a small amount of time to establish its oscillation.
Similarly, when the wind is removed, the sound does not mathematically disappear in zero seconds.
So I add an ADSR envelope.
For a Principal I might begin with something approximately like:
Attack: 10-30 ms
Decay: very short or minimal
Sustain: 100%
Release: perhaps 50-150 ms
Then I adjust it by ear.
The gate from MIDI-CV goes to the ADSR.
The ADSR controls the VCA.
So our patch has now become:
MIDI-CV pitch -> VCO
VCO -> Mixer -> VCF -> VCA
MIDI-CV gate -> ADSR -> VCA control
We now have a polyphonic organ voice that starts and stops rather more naturally.
Step Five — The Tiny Detail That Makes a Huge Difference: Chiff
At this point the sound may be recognisably organ-like.
But there is still something missing.
Listen carefully to a real flue pipe starting.
Before the stable musical tone is completely established, there can be a very brief transient produced as air begins interacting with the mouth of the pipe.
Organists and organ builders often describe part of this attack character as chiff.
And chiff is extremely useful to us because a tiny amount can transform the realism of a synthetic pipe.
The nice thing is that we can create it ourselves.
Add a Noise generator.
Filter the noise so it does not sound like broadband television static.
Then pass it through another VCA controlled by a very short envelope.
Something like:
Attack: virtually zero
Decay: 20-60 ms
Sustain: zero
Release: very short
Then mix this very quietly with the main Principal tone.
The emphasis is on very quietly.
If I can obviously hear:
psshhh — psshhh — psshhh
on every note, I have overdone it.
Ideally I should notice something missing when I remove it rather than something obvious when I add it.
That is often the difference between sound design and special effects.
Step Six — Pipes Are Not Perfect Oscillators
Our virtual VCO has another problem.
It is too good.
An electronic oscillator can sit relentlessly on its mathematical frequency.
A real pipe is being excited by moving air.
Wind pressure changes slightly.
Air moves inside the instrument.
Temperature changes.
Other pipes speak.
The acoustic environment contributes.
None of those effects needs to be dramatic, but absolute digital perfection can itself sound artificial.
So I can introduce a tiny amount of slow pitch modulation.
For example:
LFO -> attenuator -> very small pitch modulation
Perhaps start with a modulation rate somewhere around 0.2-0.5 Hz and turn the modulation depth down until it is barely perceptible.
We are talking about a few cents of pitch variation, not theatrical vibrato.
If you can hear the note obviously wobbling, it has gone too far.
Remove the modulation.
Listen.
Add a tiny amount.
Listen again.
That process is becoming a recurring theme in this experiment.
The realistic version is frequently the one where the effect is only just present.
Step Seven — One Pipe Is Not in a Vacuum
We now have something resembling an individual Principal pipe.
But a pipe organ is almost never experienced with your ear centimetres away from a pipe mouth.
You hear:
pipe + organ case + building + reflections + reverberation
This is tremendously important.
The same synthetic pipe played completely dry can sound surprisingly unimpressive.
Put it into a believable acoustic and suddenly the brain starts accepting it as an organ.
For Rack users wanting a free reverb option, Valley Plateau is one widely used module available in the VCV Library.
For this exercise I would avoid turning the reverb into an enormous cathedral just because I can.
A Principal should remain articulate.
Instead, I would begin with a modest church acoustic and adjust:
- reverb time;
- pre-delay;
- high-frequency damping;
- wet/dry balance.
Then play some chords.
Listen particularly to the release.
Suddenly our signal path has become:
Pergamon
-> MIDI-CV
-> VCO
-> waveform mixer
-> VCF
-> VCA
-> main envelope
plus
Noise
-> filtered chiff
-> short envelope
then
Principal + chiff
-> reverb
-> Pergamon/audio system
That is a surprisingly sophisticated musical instrument created from a handful of virtual modules.
And VCV Rack Can Do This Polyphonically
This is one reason VCV Rack is so useful for the project.
Many of VCV's standard synthesis modules are polyphonic, and Rack's polyphonic patch cables can carry up to 16 channels. A MIDI-CV module can therefore allocate incoming notes across multiple voices without me manually creating sixteen separate oscillators, sixteen filters and sixteen amplifiers.
From the player's point of view I simply play a chord.
Behind the scenes VCV Rack effectively creates a separate synthesis path for each note.
That makes the software much more practical for organ work than the classic image of a modular synthesizer producing one note at a time might suggest.
There Are Two Interesting Ways to Integrate Rack with the Pergamon
The simplest concept is to treat VCV Rack as an external synthesizer.
The Pergamon sends MIDI.
Rack creates the sound.
The resulting audio comes back through an audio interface or suitable audio routing.
But there is another intriguing possibility.
The Pergamon OAX system incorporates a VST3 host, with WERSI specifying support for multiple VST instruments and effects with configurable signal paths. VCV Rack Pro can itself operate as a VST3 instrument/effect inside a DAW or compatible VST host.
That potentially makes the relationship even closer.
Rather than:
organ + separate computer synthesizer
we move towards:
Pergamon + modular synthesis environment
Compatibility, installation method and system configuration obviously need to be checked for a particular OAX setup before changing anything, but conceptually this is one of the most exciting aspects of an open software-based instrument.
Why Not Just Use a Sample of a Principal Pipe?
That is a perfectly reasonable question.
After all, the Pergamon already contains excellent organ sounds, and sampled organs can reproduce real pipes extremely convincingly.
If my only objective were:
I need a good Principal 8′
then synthesizing one from scratch would be a rather roundabout way of getting there.
But that misses the point of the experiment.
I want to understand why it sounds like a Principal.
With a sample, I hear the answer.
With synthesis, I have to discover the answer.
What happens when I remove the upper harmonics?
What happens when chiff becomes longer?
What happens when the attack becomes slower?
What happens when I remove the slight instability?
What happens if the filter tracks the keyboard badly?
What makes a Principal different from a flute?
That is enormously valuable.
From One Principal to an Entire Organ
This first Principal 8′ is only the beginning.
Once the basic architecture works, there is no reason to stop there.
Change the harmonic balance and envelope and I can begin experimenting with a Gedackt 8′.
Create a brighter spectrum and move an octave upwards and I can investigate a Principal 4′.
Then perhaps:
Octave 2′
Quint 2 2/3′
Mixtures
Flutes
Strings
Reeds
A synthetic reed presents a completely different challenge from a flue pipe.
Then things become even more interesting.
Instead of reproducing an organ that already exists, I could design stops that no organ builder could realistically construct.
A Principal could gradually morph into a wavetable as aftertouch increases.
A mixture could respond to a modulation wheel.
Pipe tone could be combined with granular textures.
A 32′ synthetic bass could exist without requiring a pipe approximately the height of a house.
At that point the Pergamon stops being simply a digital recreation of established instruments.
It becomes an instrument-design platform.
The Pergamon Is Almost the Perfect Controller for This
This is perhaps what fascinates me most about the project.
A modular synthesizer is normally controlled from a relatively modest keyboard surrounded by knobs.
The Pergamon gives me something completely different.
Three manuals.
Pedals.
Drawbars.
Expression pedals.
Buttons.
Preset controls.
A large playing surface.
Suddenly I have the physical architecture of a major organ connected to a modular synthesizer in which almost anything can control almost anything else.
The drawbars need not necessarily control drawbar organ harmonics.
One could become:
chiff amount
another:
harmonic brightness
another:
wind instability
another:
reverb
another:
filter tracking
and another:
mixture strength.
Then the instrument becomes performable rather than merely programmable.
That is a very important distinction.
A Useful First Patch
For anyone wanting to reproduce this experiment, I would deliberately keep the first version simple.
Core modules
VCV MIDI-CV
Receives the notes from the Pergamon.
VCV VCO
Creates the basic waveform.
VCV Mix
Blends triangle and saw components.
VCV VCF
Shapes the harmonic spectrum.
VCV ADSR
Creates the main pipe envelope.
VCV VCA
Controls the note amplitude.
VCV Noise
Provides the raw material for chiff.
Second ADSR/VCA
Creates the brief chiff transient.
VCV LFO
Adds extremely gentle instability.
Reverb
Places the synthetic pipe in an acoustic.
Audio
Returns the completed sound to the audio system.
VCV's current free-module collection includes the main synthesis building blocks required for this patch — oscillator, filter, amplifier, ADSR, mixer, LFO and noise — while MIDI and audio interfacing are provided by Rack's Core modules.
Do not start by installing another hundred modules.
Make one convincing pipe first.
Then improve it.
A More Advanced Experiment — Stop Making Every Pipe Identical
There is one obvious weakness remaining.
If every note uses exactly the same oscillator shape, envelope and instability, then we have effectively constructed the same virtual pipe 61 or 76 times at different pitches.
A real rank is more complicated.
This suggests a fascinating second-stage experiment.
Gradually alter parameters according to keyboard position.
For example:
lower pipes could have slightly different harmonic content;
upper pipes could speak fractionally faster;
chiff could vary through the compass;
filter behaviour could change;
minute random tuning differences could be introduced.
One could even deliberately create tiny pipe-to-pipe variations.
At that point our synthesizer is moving towards physical behaviour modelling rather than simply subtractive synthesis.
And that deserves an article of its own.
What I Have Learned from Building Just One Stop
One of the things I like about modular synthesis is that it makes vague musical descriptions much more precise.
We can say:
"That sounds too synthetic."
But why?
Perhaps:
the oscillator contains too many upper harmonics;
the envelope starts too perfectly;
every note is exactly in tune;
the chiff is missing;
the filter does not track pitch;
the release is instantaneous;
or the sound has no acoustic environment.
Each objection becomes something we can investigate.
And that is why I find projects like this much more interesting than simply downloading another preset library.
I am not merely acquiring another sound.
I am learning why the sound works.
Conclusion — An Old Instrument Meets a Very New One
There is something rather pleasing about using one of the newest forms of musical software to investigate one of the oldest forms of keyboard instrument.
The principle of a flue pipe is centuries old.
VCV Rack represents a thoroughly modern approach in which virtual oscillators, filters, envelopes, control voltages and processors can be connected on screen in almost any arrangement imaginable.
And sitting in front of all of it is the WERSI Pergamon — looking unmistakably like an organ, but increasingly behaving like an entire electronic music laboratory.
My first objective is modest:
Make one convincing Principal 8′.
But once that pipe exists, the interesting question becomes:
What should I build next?
Perhaps a Gedackt.
Perhaps a Celeste.
Perhaps a Trompette.
Or perhaps something that has never existed in a pipe organ at all.
Because once an organ console is connected to a modular synthesizer, we are no longer restricted to selecting stops designed by somebody else.
We can start designing the instrument itself.
No comments:
Post a Comment