An appeal for a greater understanding of rudimentary modular functions
Making music with an analogue modular synthesizer is about the flow of electricity. Messing with that flow is fun and can lead to new sounds and structures. We use patch cables to redirect electrons around our system; patching is a form of circuit modification, a hack. In the same sense that we customize our synthesizer by the choice of modules we populate it with, we can also re-purpose and re-imagine them once they are installed.
Understanding the fundamentals helps us get the most from our system and can lead to musical results we might not otherwise have considered. It’s a philosophy borne out in the modular systems of Serge, Doepfer and others, which places the emphasis on building blocks rather than ready-mades. It’s a low-level approach, the idea that one can derive complexity and variation from a handful of simple, interacting functions.
At its most basic, the analogue modular synthesizer is all about voltages — a sound source can be given a role where it is not directly audible but still has an impact. The notion that “anything can be anything” might be exhilarating or bewildering, depending on your viewpoint or musical needs. The live performer, for example, is unlikely to focus on how clever his or her patch is and, conversely, their audience is unlikely to care. What counts are the musical results.
So, what is it about making music with a modular synthesizer that is different?
Many Strings, One Bow
Let’s have a look at some examples that highlight both the micro and macro levels of making music with a modular.
The Serge Universal Slope Generator is the quintessential multi-functional module. It’s the embodiment of what Serge users refer to as “patch programmability”. This circuit can be a sound or voltage source, or a sound or voltage processor. Amongst users, it is often referred to as the “Swiss Army Knife” and that’s why you’ll find it or its modern progeny in almost every modular musician’s system, regardless of format. The importance of this design cannot be overstated; if you understand the Universal Slope Generator, you will understand the power and appeal of making music with modular synthesizers. Serge users know it as the DSG and that’s the abbreviation I’ll use here. But you’ll also find variations of this circuit lurking behind the Bananalogue VCS or Make Noise Maths. 1[1. For a full description of its circuitry and functions, read Timothy Stinchcombe’s authoritative article, “The Serge VCS: How it works,” also published in this issue of eContact!] For the purposes of this discussion, we’re interested in two aspects of the DSG: its circuitry and its universal nature. The amplifiers, resistors, transistors, etc. allow us to do many useful things like generate oscillations, envelopes or portamento. But the fact that it is a jack-of-all-trades also means that there are trade-offs in precision. This is in the nature of analogue instruments and, as such, is something that I believe should be embraced.
To understand why the DSG is so versatile, take a look at Figure 1. It represents the result of an extravagant patch exercise, my attempt to break down a slope generator into its component parts. I got my hints from Ray Wilson’s Skew LFO and an envelope design described by Barry Klein. 2[2. Ray Wilson, “Low Frequency Oscillator Variable Skew,” updated 24 December 2002. Klein’s envelope design is analysed in René Schmitz’s article, “Discrete ADSR.”] It took seven modules to approximate the function of a Bananalogue / Serge VCS. Double that number if you want to clone your Make Noise Maths!
What’s interesting about this block diagramme are the components that make up the whole. Mixers, slew limiters, comparators — they are all modules you may already have in your system, either as standalones or hidden under another name (e.g., a clock divider is a chain of flip-flops). What makes the DSG useful is that all the functions that are integral to its own internal workings are also made accessible to the musician on the front panel.
When the DSG’s envelope has run its course, a gate signal is emitted from its “end” output. Internally, this is used to re-trigger the envelope so that it cycles and provides a periodic waveform. This in itself is pretty canny and suggests that we can use a comparator to get other, non-looping envelopes to cycle. But let’s leave that aside and re-imagine another role for the end out.
I mentioned silent signals, voltages that inform but are themselves not heard. Ignore the envelope now and focus rather on the effect it has on the end out, namely its timing. Set to non-zero times, the envelope acts as a middleman to delay incoming triggers or gates, for example to stall the onset of vibrato modulation to a VCO. At minimum times, the end out changes state almost instantaneously and functions as a binary logic inverter — useful if you want to flip a drum hit from off-beat to on-beat, for example. Both these functions are by-products of the circuit’s intended use.
Thanks to documentation by Rich Gold, Darrel Johansen, Marina LaPalma and others 3[3. A number of “Documents, manuals and catalogs” are available in Ken Stone’s Serge Modular document archive.], these techniques have passed into modular folklore and the versatility of the DSG itself has become the stuff of legend. But the lesson is still valid; it’s not just about getting the most from the module but also about understanding how things work. An appreciation of these function blocks can lead to the revelation that there are many ways of reaching the same goal and that unconventional solutions can lead to unexpected results.
Some Random Shit
Making music is as much about listening as it is about playing. When not used as the primary driver in a patch, random modules offer the modular musician a memory and a means of responding in a defined or arbitrary fashion. Overused, they can be clichéd, but their deployment highlights a key technique in making music with modulars: feedback.
In terms of their basic design, there are two sorts of random voltage generator: those which directly sample and hold a varying input signal and those which use shift registers to carry a string of processed data. Subtlety and precision may be derived by sorting the input or filtering the output. If this information is used to address the steps of a sequencer, the product can be further tailored, for example to output a selection of only three notes or voltages.
Let’s first look at the simpler of the two: the sample-and-hold-based random voltage generator. Fed with noise and pulsed into action, it generates an effect that, when used to control the frequency of an oscillator or filter, we are all familiar with: bleeping robots or bubbling mud. The circuit is simple: an electronic component called a capacitor acts as a rechargeable battery. Depending on how the circuit is configured, when triggered it outputs the last voltage it was fed.
The reason the result sounds random is due to the character of the input signal. Sampling white noise, such as the hiss of an out-of-tune radio, gives us a good distribution of voltages. If we were to sample the linear ramp of a sawtooth waveform at a regular, multiple frequency of the source, the result would be a rising “staircase” sequence of voltages, i.e. not random at all.
Don Buchla’s Source of Uncertainty (SOU) module 4[4. A description of this and other modules can be consulted in the “200 Series Catalogue of Components.”] refines this approach by using a string of sample and holds known as a “shift register”. The module encodes, holds and decodes binary information as part of its random voltage generation. 5[5. For a detailed description of how it works, see David Brown’s SOU clone build page.] Shift registers are the electronic equivalent of singing in a round or canon. In the SOU, instead of notes, digital data is passed from one stage to the next. Unlike the sample and hold, which offers a direct correlation between what was input and what is output, the stages of shift register are filled with digital data, zeros and ones. At any given time, this code is interpreted by a digital-to-analogue converter and made available to the musician.
This is an important distinction because it means the same content can be interpreted and output in a different form depending on the de-coding used, for example semi-tone or octave steps. Because the shift register has a longer term memory, it also makes other sophisticated tricks possible. The Grant Richter-designed Noisering module develops the theme by allowing feedback of the data — repeatable randomness.
Feedback is a key part of patching a modular synthesizer. We use it to generate non-linearities in both audio and control voltages. But we also use feedback to create inter-dependencies. A classic patch demonstrating this is the “tail-chasing configuration” described by Allen Strange in his book Electronic Music: Systems, Techniques and Controls and adapted by Doepfer in the manual to their version of the Buchla module.
The patch describes a loop: the module driving the generation of random events is itself influenced by the random events it has helped to generate, and that affects what the next set of random events might be. Depending on which parameters one choses to monitor, evaluate and modify, this recursion can have a subtle or dramatic effect on the patch.
This is what it is all about. When we patch, we build eco-systems, where the twist of a single knob can change the entire landscape. As pleasing to the ear as it may be, to focus on the sound alone — the rasp of a resonant Moog filter or the organic damping of a Buchla Low Pass Gate — is to miss the point of the modular synthesizer.
Cut the Crap
Every musician needs to know his or her instrument, so why should the modular synthesizer be any different? Maybe its own strengths and current popularity are even making mastering the modular more difficult than it needs to be.
Mystique surrounds these instruments; they can seem more than they actually are. The workflow is not always immediately obvious and it can be easy to get lost in a tangle of cables or blinded by blinking lights. One way to navigate a patch is to remember that it is merely an expression of cause and effect. A grounding in the basics can also help.
The new modular user cannot be blamed for being bewildered. Designers and manufacturers, now under boom-time pressure to deliver and sell the next hit module, are partly responsible if they obscure what it is that their instrument actually does. Unorthodox nomenclature can be fun 5[5. Although sometimes bordering on the cryptic, the design and terminology used on modules can also be the source of creative inspiration, as Chelesa Bruno (Eden Grey) explains in “An Artist’s Approach to the Modular Synthesizer in Experimental Electronic Music Composition and Performance,” also published in this issue of eContact!], but at a certain level of abstraction the user is forced to focus on individual modules rather than their functions. Gear lust, rather than knowledge, grows.
The myth of the original multi-functional module, the DSG, is also partly to blame. Its circuit is some 35 years old but still informs users and manufacturers. Modules which do only one job are now considered boring; “complex” is a desirable attribute. The designer Serge Tcherepnin re-fashioned his basic circuit to provide different modules with different functions. Today, the reverse is true with single digital modules including an absurd range of unrelated functions — the proverbial kitchen sink. But the DSG’s versatility is the circuit itself. Its various uses are merely a matter of patch perspective and imagination. It needs no switchable modes, “Easter eggs” nor alternative firmware; it is what it is. That’s not to reject microcontroller-based solutions. There are some jobs that computers do better and more efficiently, like quantizing a voltage to yield certain notes and scales. But a point is reached where we deny the instrument its own voice. By adding frets to a violin we enforce our musical demands on the instrument rather than playing the music of the instrument. The problem manifests itself in the planning of modular systems when driven by the desire for cool, instant gratification, all-in-one solutions or plug-ins.
The distillation of functions to a sophisticated product can, of course, make sense, for both ergonomic and economic reasons. But often compromises are made which prevent a re-purposing of functions; in other words, value is lost. For example, the Buchla 259 “complex VCO” 6[6. The construction of this module is described in David Brown’s article, “Buchla 259 Complex Wave Generator Module.”] has modulation and timbral capabilities built in, making it a boon for the live performer. However, its wave-folder or the phase discriminator used for sync and tracking are not available for external use. It’s a mature product, a well-considered patch-in-a-box, but it’s also a closed shop.
For me, it’s this packaged, directive approach that marks the real difference in design paradigm. The historical schism in modular synthesis was not between the East and West Coast, nor tied to the use of filters or frequency modulation in sound creation. Now, with more modules, analogue and digital, focussed on the end result, the need to get to know the basics is more important that ever. It needn’t be hard work — the modular synthesizer invites playful experimentation. But reflection on what’s happening at the singular level can give us a much better appreciation of what’s happening to the whole.