Well, today, we are starting our first series of educative posts about synthesis! Specially subtractive synthesis.
Why is it called subtractive synthesis? Very easy: it implies the use of a filter to cutoff or emphasise some of the harmonic content of the raw sound in order to shape it.
In this kind of synthesis, we’ve got three essential modules or parts for creating the sound: the oscillator, the filter and the amplifier. Anything with these three blocks in that specific order is a subtractive synth (miscalled by many as analog synth)!

Now, let’s get started with the very first of these basic components: the oscillator.
As you might think, the oscillator is what actually produces pitch. The oscillator in a synth does exactly the same as a string in a guitar: oscillate in an audible (or not… more on that later) range, say 20hz to 20khz. It’s the sound source. Unlike a guitar string, an oscillator is always doing its thing: oscillate, we’ll see more on this when we talk about the amp. An oscillator can output diverse wave forms, typically those are:
*Sine
*Triangle
*Sawtooth
*Ramp
*Pulse
*Noise

These are ordered from less harmonic content to more harmonic content.
A sine wave has no harmonic content at all, it’s just a pure tone, and at the other end, noise, has no fundamental tone. Sine wave, as it has no harmonic content, is not a suitable wave for subtractive synths. Why? it just has nothing to filter. Remember, no harmonic content = nothing to filter = useless for subtractive synthesis.

The triangle wave, sounds dark. It only has odd harmonics, which roll off pretty fast, and that’s why it sounds so dark.

Then, Sawtooth wave and Ramp or inverse sawtooth wave, who sonically have no differences at all, are brighter sounding, even buzzy, it reminds me very much to the sound of the Farfisa Compact… Anyway, all of this happens because this waves have much more harmonic content than the triangle, and, because it has all the harmonics, odd and even.

Moving on, we find the pulses, and here, I’m gonna talk about the pulse width. A square wave, is a pulse with 50% width. It has only odd harmonics, but, they don’t roll off as fast as what happens in the triangle. The sound produced by the different pulse widths differ pretty much one from another, from silence (0% pulse width) to nasal, to kind of plucked strings. Most oscillator designs provide some way to change in real time the pulse width when outputting a pulse wave.

Some designs, like the Juno 6, 60 and 160 series, are able to output all the wave shapes at once or mix them in any combination. Other designs like some Moog synths, allow you to output anything in between two different wave shapes, for example, anything in between a triangle and a sawtooth.

Some modern designs like the Arturia Minibrute and Microbrute have some controls to vary wave shapes like triangle and sawtooth, by folding them. Other designs, like the one found in the Waldorf Rocket, can output duplicated and detuned sawtooths from just one oscillator. In fact, this was first introduced by Roland in the early 90s in some of their digital synths. This wave shape is known as Supersaw, and it can be heard as lead in numerous trance hits, specially during the 90s.

Most synths implement two or three oscillators in their design, letting the user detune one against the other. This creates the typical fat sound heard in basses and leads. Many of this multiple oscillator designs can be hard synced, making one oscillator master, and another, slave. This means that when you detune the slave oscillator against the master oscillator you will hear for a while the detuned harmonics from the slave while it is forced to be in sync with the master oscillator. Very complex and scream-ish sounds can be achieved detuning the slave oscillator in sync mode. This can be typically heard in many Prophet 5 leads.

Now, a very different matter is how is the oscillator pitch being produced and controlled. Remember this guy is always oscillating. So, in the late 60s, Dr. Bob Moog came out with the first (usable) way to solve this issue: the Voltage Controlled Oscillator or VCO. This, as its name suggest, provides means to control the oscillator’s pitch with very low voltage signals. In fact, this signals are so low that the most adopted standard provides 1 volt per octave. This means that every 1/12 volt, there’s a change in 1 semitone in pitch.
He also came out that the best solution to control and determine how voltage is being sent to the VCO, this was to use an organ type keyboard to send different voltages from each key. So, in this case, in a range of 1 octave, lets say from C2 to C3, we have a difference of 1 volt between C2 and C3, so, if C2 is providing 1,5 volts, C3 will be providing 2,5 volts. Easy. That also means that C#2 will provide 1,500833 volts exactly, in this case, and so on.

This was the standard until the 80s and microprocessors. At that time, VCOs, were a little bit unstable, cause, any change in temperature, will make the pitch drift. So, the now affordable microprocessors, did exactly this (and still do), but in the digital domain. These is called DCO (digital controlled oscillator). Some great synths like the Roland Juno (6, 60 and 160) used DCO, and despite the oscillators are digitally controlled, they still are analogue synths.

As microprocessors came even cheaper, oscillators evolved for being digitally controlled to being entirely digital, meaning that the waves are being created in the digital domain.

DCOs and Digital Oscillators produces pitch that are so stable, that they even sound cold or thin, lifeless.

That’s why, in modern designs, like some DCO synths produced by Dave Smith’s Instruments, have some function to make the pitch a little less unstable, which is perceived as a warmer, more organic sound. Remember, VAs and soft synths, they all use digital oscillators (as if they had a choice…). Analog synths can use DCOs, but mostly use VCOs. Each has its pros and cons. For example, digital oscillators, are capable of producing really complex and bizarre wave shapes that are just impossible for a VCO or DCO for that matter. But, yet, VCOs, are warmer, and in general are perceived as more organic than digital.

Oh, I almost forget: Noise. Noise is really complex, because it has all frequencies at the same level, this means noise has no fundamental pitch, in fact, you can’t tune noise. Obviously, noise is not used to produce melodies or basslines, but, it’s very useful for creating percussion sounds like hi hats or snare drums, or for creating sound effects like wind, seashore, explosions, etc. Most synths that incorporate noise, have a dedicated oscillator for noise. The only synth that I can recall that incorporates noise as a waveform for an oscillator, is the Korg Ms-20, but, I’m sure there are some more out there.

Well, I believe that this, pretty much covers everything you need to know about oscillators. If you have any questions, you can add them below in the comments and I’ll try to answer (if I know the answer, of course). Thanks for reading, and if you like what we do, share it with the rest of the world!!
Cheers!