When it comes to synth sounds, we often speak about them being more or less complex. This refers to how much harmonic content (and inharmonic content, at times) the sound has – a simple sound has fewer harmonics (or perhaps none) and a complex sound has more. To put it another way, the more complex a sound, the more harmonic content it has. Without getting deep into the weeds of the harmonic series/harmonic theory, the basic concept is that any sound can be broken down into a collection of sine waves (the simplest wave – no harmonics beyond its fundamental frequency) related to the base frequency (i.e. the pitch) of the sound. (This is a whole fascinating topic and if you want to learn more, there are some good resources out there on it, but it’s beyond the scope of this series – frankly, knowing in detail how it works doesn’t actually help all that much in the practical matter of creating sounds.)
Given that understanding, there are two basic ways to approach synthesis. The first is taking relatively complex sounds (traditionally sawtooth and square waves; later, wavetables, samples, etc.) and then removing harmonics to get to the final timbre. This is generally known as subtractive synthesis, but you might also hear it called East Coast synthesis (I cover the reasons for this a bit later in this lesson) – we’ll refer to it as subtractive synthesis.
The second approach is to start with relatively simple waveforms, usually sines or triangles, and add harmonics thru various means, like wavefolders and frequency modulation (FM). Confusingly this isn’t necessarily known as additive synthesis – additive synthesis is a thing, but it typically refers to a specific approach to adding harmonics via stacks of sine waves at varying amplitudes and ratios to get the desired sound.
Finally, it’s important to understand that nearly all synths these days are hybrid synths – that is to say, they incorporate multiple different types of synthesis approaches in one synth. More specifically, nearly all synths these days use a hybrid subtractive approach – however they generate sounds, they typically all have a filter stage (or multiple filter stages!) as a key sound-shaping tool. Therefore subtractive synthesis is the first, and arguably most important, approach we will cover.
Core Techniques
These four techniques – subtractive, wavetable, sample playback and FM – are the most common forms of synthesis you’ll encounter. Once you are familiar with these techniques, you will be able to approach the basic to intermediate programming of 90% of all synthesizers with confidence.
Subtractive: This is the most common form of synthesis/most common technique you will encounter. Subtractive synthesis refers to the use of filters to (and related tools like EQs or Low Pass Gates) selectively remove harmonics from complex waveforms. Apart from a small handful of early digital synths, nearly every synthesizer you will ever use or see will have some kind of subtractive synthesis!
Famous subtractive synths include vintage synths such as the MiniMoog, Prophet 5, Juno 6/60/106 family; modern synths such as the MiniLogue, PolyBrute and BassStation 2; and many software synths, including free options such as the AudioKit Synth One iPad synth. (DIgital subtractive synths are frequently known as VA or “virtual analog synths,” since they aim to digitally reproduce the classic subtractive approach pioneered in analog synths.)
Wavetable: A method of creating complex waveforms by means of storing segments of the waveform in a table (like a spreadsheet, in essence), and then using various means to determine the way the table is read (for example, an envelope might control the position the table is read from). Frequently, each cell of the table comprises a single cycle waveform, and these tables can be read not only forward and backward, but vertically, randomly, etc to create novel complex waveforms from simpler elements.
Many of these synths also have tools to “smooth out” these interpolations to avoid weird discontinuities that result in harsh, distorted or excessively clicky/noisy sounds. The resulting waveforms can be extremely dynamic and variable. Further, additional synthesis methods can be applied to the waves generated in this manner, such as FM or oscillator sync, and wavetable oscillators are typically paired with a subtractive architecture for further sound sculpting.
Tables can be created via processing samples (i.e. digital recordings of sound), purely digitally generated waveform segments, or even drawing the individual segments in a visual editor. Closely related to sample playback, in that a sample is essentially a static wavetable that can only be read backward or forward.
This method of synthesis can be found in classic synths such as the PPG Wave or Waldorf MicroWave; modern hardware synths such as the Novation Peak, ASM Hydrasynth and the Modal Argon; and software synths such as Xfer Serum, Arturia Pigments, and Native Instruments Massive.
Sample Playback: A method in which digital recordings are used to generate the initial sound. These samples are typically recordings of instruments such as piano, guitars and strings, synthesized sounds, and simple synth waveforms. Sample-based oscillators are generally paired with a subtractive architecture, and some also offer additional synthesis methods such as oscillator sync or ring modulation.
Sample-based synths can sound very “realistic” but also tend to produce relatively static sounds, since the samples are fixed recordings. Early sample-based synths in particular offered extremely barebones methods of altering the core sounds, leading to “realistic but lifeless” sounding patches.
This method of synthesis can be found in classic synths such as the Korg M1 and Roland JD800; modern hardware synths such as the Korg Wavestate and samplers such as the Roland SP404 or AKAI MPC; and software plugins such as Pigments or any DAW’s sampler plugin.
FM: A method of generating complex waveforms via audio-speed frequency modulation (i.e. vibrato, essentially) of simpler waveforms. Typically, in this method of synthesis, simple sine wave oscillators are stacked together in various configurations to modulate each other and create complex waveforms. It is possible to use more complex waveforms as a starting point, and to further sculpt the resulting sounds via additional methods, especially via filters (subtractive); early FM synths were usually strictly sine-wave FM but later synths added more complex waves and elements like filters for easier, and more flexible, sound creation.
For technical reasons, FM is frequently implemented via phase modulation instead of actual frequency modulation; the results are substantially identical but much less computationally intensive. For all intents and purposes this is an immaterial difference, mostly of interest as technical trivia.
This method of synthesis can be found in classic synths such as the Yamaha DX7 and TX81Z; in modern hardware synths such as the Elektron Digitone and Korg opsix; and in software synths such as Ableton’s Operator and Native Instruments FM8.
Less Common Techniques
These three techniques – granular, additive and physical modeling – are less popular but still somewhat common approaches to synthesis. Familiarizing yourself with them is less critical, but still worthwhile, especially if the sounds created via these techniques are aesthetically appealing to you. And, of course, these methods may become more common/popular, and thus critical to understand, in the future.
Granular: Granular synthesis is a method of creating sounds by manipulating samples: specifically by splitting samples up into multiple tiny sub samples called “grains” which are then played simultaneously, with some degree of control over pitch, direction of playback, amplitude, and other factors. Larger grains can produce rhythmic, stuttering or delay-like effects, while smaller grain sizes tend to produce susurrating “clouds” of sound that maintain some character of the original sample. This method is also frequently used in time-stretching, and as an effect, to produce reverb and delay-like complements to a sound. A distinctive methodology that is hard to describe but easy to recognize.
This method of synthesis is used in modern hardware such as the Tasty Chips GR-1, the famous Eurorack module Clouds and as an option on some samplers and sample-playback capable synths such as the Waldorf Quantum. In software, dedicated plug ins such as Audio Damage’s Quanta/Quanta 2 exist, and granular options are often found in soft synths with sample playback capability, such as Arturia’s Pigments.
Additive: Additive synths create sounds by manipulating multiple sine wave “partials” in the time, frequency and amplitude domains. In theory, given enough partials and enough control over them, any sound can be recreated perfectly via additive synthesis. In practice, this is immensely complex and effectively impossible, though it’s definitely possible to produce plenty of interesting sounds via this methodology. Some newer, high-end additive synths can resynthesize sampled sounds into additive patches.
Vintage additive synths include the Kawai K5 and K5000 and NED Synclavier II. Modern hardware examples are rare, but there are a few, such as the Differential Audio EOSYNTH. In software, there are both dedicated additive synths such as Razor and Loom, as well as synths that offer additive options, such as Pigments. (Technically, all FM synths can do a limited form of additive synthesis as well, since they offer multiple sine wave oscillators with a relatively high degree of individual control.)
Physical Modeling: Physical modeling is a synth category of its own, encompassing a variety of methodologies to somewhat realistically model the mechanics of actual physical instruments. This includes such techniques as Karplus-Strong Synthesis, which uses very short delay lines and short noise bursts to recreate a plucked string, and digital waveguide synthesis techniques, which use delays, filters and nonlinear elements to model a physical object such as a drum membrane or a tube.
A typical physical modeling synth has tools like exciters (initial sound source, roughly equivalent to an oscillator; may be noise, very short samples, or other sources depending on the model) and resonators (roughly the equivalent of filters, but used differently), along with some familiar control elements such as envelopes, LFOs, etc.
Vintage physical modeling synths include Yamaha’s VL1 and Korg Prophecy; modern hardware examples can be found mostly in Eurorack, such as the famous Rings module from Mutable Instruments, and as an option in a handful of high-end synths like the Waldorf Quantum and Iridium. In software, a multitude of options exists, including Reason’s Objekt and Baby Audio Atoms.
Unusual, Rare, Obsolete and “Fake” Techniques
This class includes all remaining techniques and approaches. Some are rare because they have limited use. Some of these are obsolete, at least in the sense that they are not being commercially pursued in any serious way. Some are “fake,” in the sense that they are less a definable method of synthesis and more of a marketing term or technique previously used to sell synthesizers.
East Coast/West Coast synthesis: In the early days of commercially available synthesis technology (late 1960s/early 1970s), these terms were used to describe the two broad approaches to synthesis available at the time.
East Coast Synthesis, so-called because it was associated with Robert Moog’s synthesizers, which were made on the East Coast of the US, was the now-familiar subtractive approach – square/saw wave oscillators, running thru ladder filters and voltage-controlled amplifiers, modulated by envelopes and LFOs, usually played via a traditional keyboard.
West Coast Synthesis, so-called because it was associated with Don Buchla’s synthesizers, which were made on the West Coast of the US, relied on simple waveforms (sines and triangles, typically) that were used with techniques such as wavefolders and frequency modulation (aka FM) to add harmonic complexity and richness. These were then fed into low pass gates (a kind of combination amp and filter with a built-in decay), modulated by more complex function generators (a sort of combination envelope and LFO that’s much more flexible than either individual tool) and random sources, and often played via non traditional means such as capacitive touchplates rather than a keyboard.
In truth, both methodologies overlapped considerably, and can be used together for even more flexibility. These days, this is mostly a historical and philosophical distinction, rather than a defined set of tools or techniques.
Vector Synthesis: A technique and marketing term that describes the mixing of multiple oscillator sources under control of either a joystick or envelope. This would provide a way to animate the core oscillator sound to provide a more varied oscillator sound.
Vintage examples include the Sequential Prophet VS and Korg Wavestation. Modern hardware examples include the Korg Wavestate. Software examples include emulated versions of the previously named synths.
Phase Distortion: Phase distortion is an early digital synthesis technique related to FM/PM (frequency modulation/phase modulation) synthesis made famous by Yamaha. Phase distortion is a simpler method, using a modulating oscillator to distort a sine oscillator to produce complex waveshapes. Functionally, PD synths are programmed like, and sound similar to, subtractive synths despite the difference in the way the sounds are generated, and they have a distinctive sound that can mimic analog subtractive while also offering some nice, bright “digital” sounds.
Vintage synths that use phase distortion include the Casio CZ series, for which it was invented. Modern software synths include emulations of the CZ line such as Arturia’s CZ V and options inside synths including Pigments.
VA/Virtual Analog: Virtual analog is less a synthesis technique and more a marketing term for older, digital subtractive synths that mimic the workflow and approach of vintage analog synths. Essentially, virtual analog is another name for a digital subtractive synth, tho some virtual analog synths incorporated additional techniques such as FM and ring modulation.
Vintage hardware synths marketed as VA include the Roland JP-8000 and Yamaha AN1X. Modern examples include the Modal Cobalt 8 and Korg MicroKorg. In software, VA style synths abound, tho not always marketed as such in the last decade or so. The most obvious examples of VA software synths are the many emulations of famous synths available as plug ins, and synths like Ableton’s built-in Analog plug-in.
The Rest: There are many, many more approaches to synthesis out there, but the majority of what remains is either incredibly rare (eg existing only in academic music contexts, or perhaps in a few unusual Eurorack modules) or mostly marketing (eg Roland’s Linear Arithmetic synthesis, which was simply digital subtractive synthesis that added very short samples, mostly attack transients, that could be combined with the digital subtractive architecture – in other words, an early digital hybrid approach!).
Next: Subtractive Synthesis Basics
Prior: Intro and Philosophy