Back to Basics (Part 2)
Continuing our series for the complete newcomer to electronics in music, Steve Howell takes a look at the VCO, the synthesiser's heartbeat.
The second part of our series for the complete newcomer to the world of synthesisers.
Having looked at the basic working principles behind every analogue synthesiser, we can now safely move on to discuss individual modules in turn. And what better place to start than the foundation stone from which all analogue synths are built? I refer of course to the oscillator, and to begin with, those oscillators whose operation is governed by the principles of voltage control discussed last month.
An oscillator's purpose is simple: to provide a basic tone at a variety of different pitches. That basic tone takes the form of a variety of waveforms, whose pitch is determined primarily by the synthesiser's keyboard. Before we start going into too much detail about a VCO's output, however, perhaps a quick lesson on the physics of pitch wouldn't go amiss.
In order for human beings (and I imagine that probably covers most of you) to perceive the pitch of a sound, that sound must vibrate (or oscillate) at a rate of between 40 and 15,000 cycles per second, or as technical terminology would have it, 40Hz-15kHz. And one rule we can apply to all sources of sound is that the more cycles there are in the space of a second, the higher the sound's pitch will be perceived as being.
If you're having problems visualising how this principle might work out in real life, imagine the workings of a common or garden record player. Play a 331/3rpm disc at 45rpm and its pitch sounds higher than it ought to be, and it's much the same with synths: the faster a VCO oscillates, the higher its pitch will be. The two parameters work in proportion to each other, too, so that a doubling of frequency results in a doubling of pitch. In other words, 200Hz is an octave higher than 100, 400 is an octave higher again, and so on until the end of the Universe.
How does electronics get into all this? Well, purely and simply because applying a voltage to a VCO's input alters its resultant pitch. And one volt per octave synths - which is a lot of them - usually follow the rule that one volt of change will result in a pitch change of one octave, so that, for instance, a 1V input gives a 200Hz output, 2V gives 400Hz, 3V gives 800Hz, and so forth. See Figure 1 for a pictorial realisation of this concept (man).
Even with the help of the diagram, this theory might seem a little on the complicated side to some of you, but if it does, don't panic. For one thing, it's really the only bit of theory you have to know in connection with oscillators, and for another, it won't really harm you all that much if you don't really get the drift of it at all. Take the time to get to know it, however, and it'll gain you a greater understanding not only of synthesisers but of sound in general.
So, having gotten the heavy stuff out of the way, let's look at the VCO in detail - Figure 2 shows the layout of a typical one. You'll see that this imaginary oscillator outputs three different waveforms, selectable by means of a simple switch. It also has an input mixer that allows you to mix a variety of control sources such as the keyboard, the low frequency oscillator (LFO), envelope generator (EG), and pitch-bend levers, to name the most common elements. The mixer also has the job of mixing in a number of DC voltage sources necessary to set the basic range of the VCO and tune it to other sound sources, and because it goes about its business by adding voltages together, the mixer is known as a Summing Amplifier.
How does it work? Well, let's say you press Middle C on the synthesiser keyboard: this will send 1V to the VCO's summing amplifier. If you then adjust the Tune control to give a 1V output from the DC source, this will add another volt to the summing amplifier and the pitch will rise by an octave as a result. Move the pitch-bend wheel or lever down to its minimum position and you'll add -1V to the summing amp, so the pitch will go back down an octave.
The above is just an example that illustrates how voltages are added and subtracted in different ways to produce different pitches, but in fairness, you're unlikely to need to check voltage readings from the various controls with a multimeter every time you set up your favourite string sound, so it's not a process that needs to be retrieved from the memory cells every time you start programming your synth. On the other hand, knowing how the input stage of a VCO gets its act together will enable you to predict what results your programming manipulations are likely to have on the final sonic outcome. Which is a much more agreeable situation than aimlessly fiddling with your instrument in the hope of getting some sort of result (know just what you mean - Ed).
But enough of input stages for the moment. Let's look at what happens at the other end of a VCO's scale of operations by trying to discover what waveforms are all about. The bad news here is that I'm going to have to delve into the world of physics again to start off with, I'm afraid.
Most sounds contain a phenomenon you've probably all seen referred to in the past as Harmonics. These are actually rather important as they're what enables us to perceive the tone (or as the French, and probably the Californians, would call it, the timbre) of a sound. And as a general rule, the more harmonics there are in a sound, the brighter we perceive that sound as being.
What are harmonics? Quite simply, they're multiples of the fundamental waveform, the sine wave. In fact, as Figure 3 shows, harmonics are also sine waves, as multiples of the fundamental's frequency.
So having said all that, why have different waveforms in the first place? Well, the reasoning lies in the fact that most sounds, in addition to having more or less harmonics, also have them arranged in various different ways. For example, a trumpet has both odd and even harmonics, while a clarinet has only odd ones. And in order that the poor, unsuspecting synth player might have a bash at programming sounds that approximate as many acoustic equivalents as possible (as well as voices that have no precedent in the acoustic arena at ail, of course), synth designers have kindly seen fit to endow their produce with waveforms that possess different harmonic structures.
The most commonly found waveforms are as follows. First, the sawtooth, a waveform that comprises both odd and even harmonics, and sounds bright (and vaguely brassy) in its raw state, ie. unfiltered by any later synth module or signal processor.
The square wave, by contrast, contains only odd harmonics and is therefore useful for mellower sounds such as those of woodwind instruments. It can also come in handy for injecting 'welly' into bass and lead line sounds. Pulse waves are a little bit complicated as they don't technically constitute a classification of their own: they're usually obtained by varying the width of an existing square wave. And because pulse waves possess a continuously variable waveshape, it's impossible to say precisely what their harmonic structure is. Suffice to say that most pulse waves sound thin and nasal, and are therefore well suited to the imitation of delicate acoustic timbres such as harpsichord sounds. It's possible to vary the width of a pulse wave automatically using voltage control - this sweeps the width from one extreme to the other and is useful for chorus and flanging effects à la Gary Numan.
The triangle wave is similar to the square in that it also contains only odd harmonics: the difference is that there are less of them, and as a result of this, the waveshape's sonic output is more muted than that of a square wave and is useful as a source of flute sounds and the like.
The sine wave is totally devoid of harmonics, and is therefore the purest waveform available to mankind. In practical terms, it's not actually a particularly useful waveform, though its ability to reinforce the fundamental of a sound makes it useful for beefing up weak voices. As it happens, few synth VCOs are equipped with sine waves as they're none too easy (read 'cheap') to reproduce electronically without any distortion. If you're one of the majority without a sine option on any of your variable oscillators, you can use a filter to do the same job, but we're beginning to leap ahead of ourselves.
The waveshape symbols illustrated in Figure 4 are graphic representations of what you'd see if you connected the VCO's output to an oscilloscope. A synth from one manufacturer should produce much the same visual result as a model from any other, so these symbols have now been almost universally adopted as the standard means by which different waveforms can be identified visually. Once you know what each symbol represents, you're halfway towards achieving positive programming results.
You'd be quite blameless in thinking that what we've talked about so far this month is quite complicated, but in reality, we've only uncovered the tip of the waveform iceberg. It's all very well discussing individual waveshapes in isolation, but what happens when we start using different waveforms in combination with each other?
Consider that all mixer modules present on an analogue synthesiser - whether voltage or audio - are summing amplifiers, and are capable of adding one voltage to another to form a composite result. The waveforms generated by VCOs are little more than AC voltages oscillating at audio frequencies anyway, so they can be added together to create new voltage shapes, or in other words, new waveforms.
Well, that's the technical version. All you really need to know in practice is that by combining waveforms together, you can drastically increase the range of sonic possibilities at your disposal.
Figure 6(a) shows the effect of mixing a sawtooth with a square wave, and the relative levels of each waveform should yield a variety of different timbres. Figure 6(b) shows a similar mixing process, except that the two waveforms concerned are both of the sawtooth variety, tuned slightly apart. The result of this is that the waveforms are out of phase with each other. Depending on just how far the two oscillators are tuned apart, this arrangement is capable of providing a broad range of effects from mild flanging to heavy chorus. The fluctuation in level caused by the phase difference between the two VCOs is known as Beating, and the wider the pitch difference between the waveforms, the more rapid this beating becomes.
In addition to tuning oscillators only slightly apart for ensemble and associated beefing-up effects, it's also possible - on many two-oscillator synths - to tune them apart by a standard musical interval such as an octave or a fifth, which has the thoroughly agreeable effect of making your lone synth sound like it's twice the instrument it really is.
As I've been at pains to point out, an in-depth knowledge of how your synthesiser does everything it's capable of is by no means a prerequisite for enjoying playing and programming it. After all, you don't have to know the precise mechanism by which each moving part of a piano does its job to become a great concert pianist.
Understanding the physics behind the synthesiser's modus operandi - which is what this series is all about - should enable you to get better results out of programming in a shorter time. But one trap you should beware of falling into is letting your knowledge of the technology of synthesis dictate the sort of music you play. As with any musical instrument, what you want to achieve from a musical point of view should determine how you apply the relevant technology, not the other way around.