Learning Synthesis: VCAs

Learning Synthesis: VCAs

Developing Dynamics and Modulating Modulation

Ryan Gaston · 03/12/19

It is no secret that many traditional synthesizers' sonic identity is tied up in their oscillators and filters. The oscillators are the sonic bedrock from which the final tone is sculpted; the filters add a special touch of coloration, sometimes subtle, sometimes extreme. And while even just these elements alone are enough to provide a wide array of timbres, one would struggle to create sounds with varied dynamic profile, or sounds with timbral subtleties that evolve over time.

This is the domain of the Voltage Controlled Amplifier: a synthesis building block that carries the capacity for all sorts of nuance, from note articulation to management of modulation. VCAs are a core synthesis element, most commonly used to handle dynamics—though we'll quickly see that they are capable of much more.

Amplitude and Multiplication

Because of its other musical affiliations, the word "amplifier" when applied to synthesis can be a bit confusing. And while they are in some ways conceptually similar, amplifiers for guitar or live sound are quite different in function from VCAs. Guitar amps and pro audio power amplifiers primarily serve the purpose of greatly increasing (amplifying) a signal so that it is powerful enough to drive large speakers, which then create dramatic fluctuations in air pressure which we then perceive as loud sounds. And while loud sounds are awesome, that's not 100% what VCAs are about.

Amplitude of range A is 1V; amplitude of the waveform as a whole is 2Vpp, as demonstrated by range B. Amplitude of range A is 1V; amplitude of the waveform as a whole is 2Vpp, as demonstrated by range B.

Amplitude refers to a signal's magnitude in terms of voltage level: a line-level signal generally has a maximum amplitude range of up to around -1V to +1V (or 2V peak to peak, 2Vpp), while modular synthesizer signals can have amplitudes in the range of -10V to +10V, or 20V peak to peak. For audio signals, amplitude becomes a measurement of loudness; for modulation signals, amplitude corresponds to the intensity of the modulation effect.

The traditional definition of “amplification” can be a bit confusing in this context: despite its common use, amplification on a technical level is really all about multiplying a signal's amplitude by some value (in fact, it's not uncommon to refer to VCAs as multipliers or signal multipliers). The value by which the signal is multiplied is generally referred to as gain. Guitar amps can boast very high gain, boosting signals up to hundreds of times their initial level. Synthesizer VCAs, on the other hand, deal with relatively low gain: amplification factors are usually in the range of 0–1x the original signal level. Let's unpack the implications of gain factors this low.

Three signal multiplication examples: a signal multiplied by 2 (top); a signal multiplied by 0.5 (middle); a signal multiplied by 1 (bottom). Three signal multiplication examples: a signal multiplied by 2 (top); a signal multiplied by 0.5 (middle); a signal multiplied by 1 (bottom).

Multiplying a signal by a value of 1 results in output level identical to the input: we generally call this unity gain. Multiplying a signal by a factor of 0, however, will result in silence. And of course, multiplying by any value between 0 and 1 results in a signal with identical shape to the input, but an overall lower magnitude: multiplying by 0.5, for instance, results in a signal with half the amplitude of the input. In this way, one could think of VCAs as voltage controllable attenuators, as they do generally serve to lower voltage levels rather than increasing them. (This is, of course, not always the case—there are plenty of VCAs out there that can easily amplify beyond unity gain.)

When making music, we seldom think so precisely about these factors—instead, we more often learn intuitively how different VCAs respond to different signals, and how we might use the yielded effects in our music. As such, one does not necessarily need to think about gain factors and signal multiplication when patching or performing, but at least this much should be clear: a VCA can, in effect, act like a volume control for a sound.

But What About Voltage Control?

As with VCOs and VCFs, the most exciting and practical ramifications of the VCA are found by exploiting its potential under voltage control.

VCAs are a fairly universal part of analog synthesizers' signal paths. Often, the audio signal path originates at a VCO, travels through a filter, into a VCA, and then straight to the synth's final outputs. In this scheme, the VCO provides the basic sound—the filter then colors the basic timbre, and the VCA acts as a final dynamic control for the synth voice.

Simple envelope-controlled loudness, with the yielded waveform in the upper right and envelope shape in the bottom right. Simple envelope-controlled loudness, with the yielded waveform in the upper right and envelope shape in the bottom right.

In this common subtractive setup, the voltage control for the VCA usually comes from an envelope. The shape of the envelope itself then directly correlates to the gain of the VCA—and the constantly variable gain results in a sound whose loudness varies over time. The dynamic contour of the sound can then be controlled directly by the envelope itself: in most keyboard synthesizers, each voice has a dedicated VCA envelope meant to offer easy access to a wide variety of dynamic articulations.

Simple sine wave tremolo, with the yielded waveform in the upper right and the modulation waveform in the bottom right. Simple sine wave tremolo, with the yielded waveform in the upper right and the modulation waveform in the bottom right.

Of course, envelopes are not the only signal sources that can modulate a VCA's gain. It is common, for instance, to use LFOs to modulate a signal's level. In this case, the net result is a tremolo effect. Both the modulation wave shape and amplitude are of important. The shape determines the relative smoothness of the tremolo: sine and triangle waves produce continuous volume changes; saw waves produce repeating swells with abrupt endings; square wave produce sudden alternations between low and high volumes. The amplitude of the modulation waveform determines the overall intensity of the tremolo effect, i.e. how dramatic the variation in volume is.

Advanced Applications: Amplitude Modulation

Buchla's 144 Dual Square Wave Generator, which offered dedicated inputs for Amplitude Modulation. Buchla's 144 Dual Square Wave Generator, which offered dedicated inputs for Amplitude Modulation.

This same technique can bear very different types of sounds when pushed to extremes. Much in the way that FM can be considered an extreme variation on vibrato, Amplitude Modulation (AM) is an extreme variation on the concept of tremolo. By pushing the modulation waveform out of low frequency range and into audio rate, very extreme sounds can result: everything from bizarre distortion-like tones to clangorous inharmonic sounds to strong reinforcements of a sound's natural harmonic structure.

Much like with FM, the modulation waveshape and ratio between the modulation frequency and carrier frequency are of critical importance in AM. Sounds with simple ratios between modulation and carrier frequency (2:1, 3:1, 4:3, etc.) produce stable timbres with an identifiable pitch (or set of pitches), while more complex ratios (23:17, 532:77, etc.) result in clangorous, inharmonic sounds. The effect of waveshape is similar to its manifestation in tremolo: square AM modulation, for instance, produces fairly abrasive timbres while sines, for instance, create much smoother, cleaner sounds. All these factors work together to provide an endless range of sonic colors. This style of modulation has roots in the early stages of electronic music, and was brought directly to the surface in one of Don Buchla's first oscillators: the 144 Dual Square Wave Generator, which featured dedicated AM inputs for each of its two oscillators.

Advanced Applications: Velocity

Aside from basic articulation control, tremolo, and AM, VCAs can of course be used for a huge range of utilitarian and creative purposes.

One of the most practical extended uses of a VCA in traditional subtractive synthesis is the implementation of velocity: that is, an extended dynamic control that enables players to alter a note's volume based on how hard they press a key on a keyboard (or by how the velocity value is defined in a DAW's piano roll, etc.). Velocity is typically implemented in analog synthesizers by using chained VCAs, which can be routed in a couple of different ways.

One potential VCA-based method of implementing velocity control in a synthesizer. One potential VCA-based method of implementing velocity control in a synthesizer.

One common approach for implementing velocity in analog and modular synths involves adding a second VCA at the end of the audio path. As in our simpler subtractive synthesis examples, an envelope generator runs to the first VCA's control input, creating the sound's overall dynamic contour. The output of this VCA is then passed into a second VCA. This second VCA is controlled directly by the velocity signal (whether it originates from a keyboard, computer, or elsewhere). The net result is a sound with a definable dynamic shape which can be louder or softer based on playing velocity, all without altering the sound's internal dynamic profile.

A second VCA-based method of implementing velocity control in a synthesizer. A second VCA-based method of implementing velocity control in a synthesizer.

A second approach to implementing velocity in a modular synth involves adding a second VCA in the control signal path rather than the audio path itself. This brings to the forefront an important idea: VCAs do not necessarily have to alter the amplitude of audio signals—they can be used to alter the amplitude of modulation signals as well.

In this scenario, the typical subtractive signal path is more or less unchanged—the only modification is the insertion of a VCA between the envelope's output and the audio VCA's control input. Velocity can in this scenario be used to alter the amplitude of the envelope signal itself, in effect changing the range of loudness that the envelope yields when connected to the audio VCA. The net result here is more or less identical to the example above with two VCAs in series—as with many concepts in synthesis, there are almost countless ways to achieve any given effect.

Advanced Applications: Indices

Simple patch in which a VCA is used as an envelope-controlled Frequency Modulation Index. Simple patch in which a VCA is used as an envelope-controlled Frequency Modulation Index.

The seemingly ubiquitous synthesis mantra about "never having enough VCAs" is based in part on their potential to act as dynamic controllers for modulation depth, as with our second "velocity" example above. The VCA in these scenarios is something like a clamp or spigot that limits the amount of modulation allowed to pass through to a destination. The intensity of the modulation can be referred to in many different ways: modulation depth, modulation intensity, and modulation index are common ways of addressing this parameter. For instance, vibrato with a low modulation index between LFO and oscillator pitch is fairly subtle; vibrato with a high modulation index yields much broader change in pitch.

Diagram of the Buchla 208's Mod Oscillator section. Diagram of the Buchla 208's Mod Oscillator section.

This type of index control can of course be managed with a simple attenuator; however, when using a VCA for managing a modulation index, the modulation intensity itself can be varied via external control sources. That is to say—one could use an envelope, LFO, or other modulation source in order to dynamically control how much modulation is applied to some other signal. Some synthesists refer to this as a means of "modulating modulation," and that is fairly accurate—it is a means of applying dynamic modulation control. Whoa.

This type of management of modulation index is now fairly common, but can also be seen in older instruments from Buchla and Serge, whose design philosophies were centered around extremely flexible management of modulation signals. Several of their designs offered built-in modulate-able indices.

One of the most obvious examples in Buchla's instruments can be found in the much-loved and fabled Music Easel's voice module, the Model 208 Stored Program Sound Source. The 208, a full synth voice module, contains what is in many ways the first instance of what we now think of as a typical "complex oscillator": a dual oscillator with waveshaping, and a built-in modulation path from the first oscillator to the second. The built-in modulation paths include FM and AM from the modulation oscillator to the primary oscillator, as well as an option for ring modulation with external signals.

The 208's modulation index itself can be modulated via any signal source: great effect is often achieved by using the same envelope that opens the unit's VCA to simultaneously increase the modulation depth, making for a sound that grows to be more complex as it gets louder. This is part of the basis of the classic "Buchla bongo" patch, in which a sound's spectral content is very rich at its onset, but simplifies as the sound decays. This type of dynamic timbre control is a hallmark of "West Coast" synthesis design, and demonstrates how a modulation index can become an evocative destination for performance control. This is, of course, not unique to Buchla's instruments: similar considerations can be found in several designs by Serge Tcherepnin, creator of the Serge Modular Music System, many of whose modules are now available in Eurorack format via Random*Source.

Random*Source's Eurorack incarnation of the Serge NTO. Random*Source's Eurorack incarnation of the Serge NTO.

Serge's New Timbral Oscillator is another spectacular example of a device with a built-in modulate-able modulation index, this time specifically for linear FM. Typically, an audio source is sent into the Linear FM input and a modulation signal into the "VC" input, enabling FM whose intensity varies based on an incoming control voltage.

Of course, similar effects can be achieved through modular means: just place a VCA between any modulation source and destination and you suddenly have a dynamic control for modulation depth that can be controlled from any modulation source you desire.

An important note to add as we discuss VCAs as depth controls for modulation signals: not all VCAs are quite as well-suited for this task as others. This often has to do with whether a VCA is AC coupled or DC coupled. Coupling in this context largely has to do with whether a VCA is optimized for use with audio signals or control signals: in AC-coupled VCAs, a simple and subtle highpass filter is internally implemented in order to prevent DC offsets in audio signals, which can cause unwanted distortions and irregular behaviors during recording and sound reproduction. As such, AC-coupled VCAs are great for audio, but not so great for low-frequency modulation signals, which often lose their shape or get filtered out entirely.

DC-coupled VCAs, on the other hand, contain no such filtering element: they let all signals pass through regardless their frequency, and as such are much better-suited for modulation applications. DC-coupled VCAs will also work just fine for audio, with the one caveat that DC offset is possible with some signals...but honestly, the results of such offsets are not really disastrous and certainly do not get in the way of having a good time.

Most Eurorack VCAs are DC-coupled, for what it's worth—making them good candidates for use with basically any signal you can throw their way.

Advanced Applications: Panning & Crossfading

Simple stereo panning patch using two VCAs. Simple stereo panning patch using two VCAs.

Spatial effects are also commonly achieved with VCAs—and while there are definitely dedicated panning modules out there, these types of effects can definitely be achieved in slightly more crude forms with a couple of VCAs and some simple utilities.

In a typical stereo panning patch, one audio signal is split and sent to two VCAs at the same time. A single modulation signal is sent to both VCAs—but it is inverted before one of the VCA's control inputs, so that the sound's volume will increase in one VCA while it decreases in the other. The modulation signal can be anything: a joystick, an LFO, an envelope, a sequencer...the list goes on. Of course, separate modulation could be applied to each VCA as well, creating disproportionate changes in level between each speaker, creating a sound that bulges and recedes while moving from left to right. This is a slightly more experimental method...but it's super cool, nonetheless.

Quadraphonic panning can be achieved through a similar implementation. By using six VCAs to create three of these panning paths, very complex panning schemes can emerge. The sound is first passed into one panner, whose two outputs are each fed into their own subsequent panners. Each set of secondary panners use the same modulation signal, with a separate signal modulating the "upstream" panner. Each of these modulation signals then independently controls the sound's balance between the front & back and left & right sides of the room, respectively. A bit odd to set up—but it totally works.

Learning Synthesis: VCAs

Also worth considering: panning such as this does not have to lead directly to two (or four, or seventeen) speakers—it can instead be used to pass a signal between multiple subsequent processing paths.

A crossfading patch based on a similar technique as the stereo panning patch above. A crossfading patch based on a similar technique as the stereo panning patch above.

Crossfading operates via a similar concept. Two sound sources each pass through their own VCAs, and the VCAs' outputs are then mixed together. As before, one control source is used to modulate each VCA, one in original form and one with its shape inverted before VCA's control input. The result is a modulate-able crossfader, where incoming modulation creates an increase in one VCA's level while proportionately decreasing the other's.

Modules such as the (recently discontinued) Harvestman/Industrial Music Electronics’ Double Andore Mk2 can be of particular use in scenarios like those described here: a dual VCA/dual envelope rife with potential for complex internal modulation. While it seems to be primarily seen as a dual envelope with some added VCAs, the sheer usefulness of the pairing of functions becomes apparent when viewed through this lens: a bizarre animated panner and crossfader with a mind of its own.

Note: for these panning & crossfading patches to work correctly, use unipolar LFOs inverted around their median value: that is, if using 0–5V LFOs, they should be inverted with 2.5V as the central axis. For best results, try using with logarithmic VCAs or logarithmic control voltages with linear VCAs.

The Black & Gold Optomix, a Eurorack Lowpass Gate inspired by the Buchla 292. The Black & Gold Optomix, a Eurorack Lowpass Gate inspired by the Buchla 292.

Weird VCAs: Lowpass Gates

The world of modular synthesis being as expansive as it is, there are definitely other ways to alter signal levels. Some other common ways of achieving VCA-like effects can be found in lowpass gates and ring modulators.

Lowpass gates (LPGs), previously discussed in our article about filters, are peculiar circuits that exhibit both filter-like and VCA-like behavior. By nature of their design (usually centered around a primitive gain control using vactrols), LPGs usually have a fairly sluggish dynamic response that leads to relatively gentle attacks and ringing decays when provided short envelopes, making for fairly plucky dynamic profile reminiscent of a bongo or marimba. Because of their combination of filtering and VCA-like behavior, they offer the added side effect of having a brighter sound at higher volumes and darker sound at lower volumes—not unlike the behavior of most acoustic instruments. For more information about the inner workings of LPGs, check out our Learning Synthesis: Filters article.

LPGs were originally found in Don Buchla's instruments—namely, the Model 212 Dodecamodule, 292 Quad Lowpass Gate, and the aforementioned 208. The sound of the lowpass gate became a definitive characteristic of Buchla's instruments, now often described as strikingly "organic" or "natural"-sounding. As such, the lowpass gate has been afforded new popularity in Eurorack format, perhaps most notably with Make Noise's QMMG and Optomix, as well as Doepfer's A-101-2 and designs from Sputnik Modular, Endorphin.es, and others.

Weird VCAs: Ring Modulators

DC-coupled ring modulators can also be used as VCAs—in fact, they are conceptually very similar, with a small caveat or two. Ring modulators, like VCAs, can be thought of as signal multipliers; generally speaking, however, VCAs will only allow for multiplication by factors greater than zero, while ring modulators allow multiplication by both negative and positive factors.

In practice, this means that VCAs generally will not respond to negative voltages at their CV inputs: instead, they typically remain silent so long as negative voltages are present at their input. Ring modulators, on the other hand, invert the input signal when provided negative voltages to their modulator input. Therefore, when providing a ring modulator a unipolar voltage (like that from an envelope) to its modulator input, the performance is identical to a VCA: but when negative voltages are applied, the incoming signal comes out "upside down." The way one might think of a VCA as a Voltage Controllable Attenuator, one could similarly think of a ring mod as a Voltage Controllable Polarizer.

Note: this is why the panning patches above require unipolar modulation inverted around a positive voltage. If negative voltages were applied to VCAs' control inputs, the effect would not be panning, but a momentary muting effect when the incoming CV dipped below 0V.

Ring modulators are more commonly used as special effects for audio—usually, two different bipolar audio signals are provided to each input to create all kinds of clangorous, inharmonic effects not dissimilar to the AM techniques described above. However, when provided low-frequency modulation signals, the effect is almost indistinguishable from that of a VCA. Some companies have taken advantage of this blurring of functionality in order to create modules equally well-suited to each use: Make Noise's Moddemix, Befaco's A*B+C , and Random*Source's Serge Ring are three such devices.

Other Worthwhile Notes: Response Curve

Depending on their intended use, modular VCAs also often feature a control for response curve: that is, a control for how the signal multiplication is scaled relative to an incoming control voltage. Many VCAs offer a strictly linear relationship between incoming CV and multiplication factor—as an incoming CV steadily increases, the level of amplification increases directly in tandem. Sometimes, though, other response curves are preferable.

Linearally ascending applied VCA control, with corresponding diagrams for the linear, logarithmic, and exponentially-derived resulting signal levels. Linearally ascending applied VCA control, with corresponding diagrams for the linear, logarithmic, and exponentially-derived resulting signal levels.

Many VCAs offer either a switch or a knob that allows the user to alter a VCA's response curve from linear to exponential response. The reason for this is a bit complex and involves an excursion into psychoacoustics...but the short explanation is that the relationship between measured amplitude of a signal and perceived "loudness" is not strictly linear, itself. In fact, the relationship between amplitude and "loudness" is more or less logarithmic: very minute changes in amplitude yield fairly significant differences in perceived loudness.

Exponential VCAs can be used to help compensate for this bias in our hearing by weighting incoming voltages generally toward the lower end of the spectrum: a linear ascending voltage instead yields an exponential increase in amplification level, which balances with our logarithmic hearing bias to generate a convincingly continuous linear volume change. So that sounds a bit dizzying, but don't worry! Luckily, this isn't neurosurgery—it's just music, and the stakes really aren't so high.

The big thing to take away from this is that different response curves will yield different results: generally speaking, people prefer a relatively exponential response for audio "volume" control, and linear response for CV amplification. Depending on your modules, you might find these response curve controls on the VCAs themselves, or on the envelopes used to open them. Either way works just fine.

Also worth noting: many VCAs when set to exponential response become considerably more sensitive than in linear response mode. Mutable's Veils is one such example: with response set to exponential, these VCAs amplify well beyond unity level. This can serve basic utility functions such as line-to-Eurorack level conversion, or can simply be used as a means toward crude distortion. Many gnarly sounds can be found lurking within even the seemingly most mild-mannered VCAs, Veils included.

Other Worthwhile Notes: Common Forms

Intellijel's Quad VCA, a 4-channel VCA with all the niceties provided. Intellijel's Quad VCA, a 4-channel VCA with all the niceties provided.

Being such a basic concept, there is not a huge amount of variation in the ways that VCAs are parameterized. Most keyboard synth VCAs offer a control for envelope intensity, and only sometimes include additional controls for default VCA level or velocity. Most modular VCAs are fairly straightforward as well: they generally offer an input, output, level control, and CV input.

Some VCAs (like Intellijel's uVCA II or Doepfer's A-132-3) include built-in attenuation for the CV inputs as well as a control for default level—this feature is generally quite handy, as it allows the user to easily tune the intensity of the VCA's response to whatever CV it is provided, as well as to set a default signal level to pass through at all times. Some VCAs though, such as ALM's Tangle Quartet, offer only one knob per channel, which acts as a level control when no CV input is patched, or as an attenuator whenever a CV input is provided.

It is also fairly common in Eurorack format for multiple VCAs to be packed together into a single module and coupled with mixing capabilities. Such is the case with some of my current favorite Eurorack VCAs: Intellijel's Quad VCA and Mutable Instruments's Veils and Blinds (though strictly speaking, Blinds fits more into the ring-mod-as-VCA paradigm). These modules are all quad processors whose outputs are cascaded into tiered mixers: as such, they can all operate as four-channel mixers, three-channel mixers with an extra standalone channel, or as a dual two-input mixer. By patching from any output, that output and all preceding channels are excluded from the next channel's mixing stage, making for several complex options for signal combination. 4MS's VCA Matrix takes this idea to the extreme, providing a matrix mixer-style configuration with mutes and a dedicated VCA at each node—perfect for a stunning array of dynamic mixing and routing scenarios.

VCAs are clearly an invaluable resource for synthesis: they form the backbone of many signal routing possibilities in modular synthesis, as well as the typical dynamic structure of almost every analog synth out there. They help to tame modulation, to enable dynamic control of signal distribution, and to tell the synthesizer when to be quiet. While they may not seem exciting on the surface, they are the cornerstone around which so many possibilities are built—it is important, as with all things, to see beyond the basic realm of how-they-function and into the weirder world of what-they-can-do.