Back to Basics (Part 3)
If you're a complete newcomer to the world of synthesisers, this is the series for you. Steve Howell takes a beginner's look at filters.
Another not-to-be missed instalment in our series for the complete newcomer to synths.
Last month we saw how the Voltage Controlled Oscillators (VCOs) on an analogue synth generate a variety of waveforms at a pitch determined by the incoming voltage. Whilst this undoubtedly gives synth players a lot of tones to play around with, it's more or less essential that there's also some way of modifying the harmonics in these waveforms a bit further. Enter the Voltage Controlled Filter...
The VCF is really nothing more than a vicious tone control that lets you subtract harmonics from the 'raw' waveforms being fed into it. You'll probably remember that the reason different waveforms sound different lies in the fact that they have varying harmonic structures. For instance, a sawtooth waveform sounds bright and brassy because it has a generous helping of odd and even harmonics, while a square wave sounds hollower because it has only odd harmonics. On most analogue synthesisers, you'll find a voltage controlled lowpass filter whose job it is to subtract upper harmonics but whilst allowing the lower ones to pass through unaffected (hence the name).
Having said that, what does all this filtering actually sound like? Well, as the harmonics are removed, the sound output gradually becomes duller and more mellow. A quick glance over at Figure 1 will reveal a spectragraph: what it shows is a sawtooth wave with the response of a lowpass filter superimposed on top of it. The shaded area represents the sound that will actually be heard - all the harmonics above the cutoff point will be attenuated, or for the more sadistically-inclined, cut off. Most filters have a control by the name of Cutoff Frequency, which lets you move that point to wherever your fancy takes you which, in turn, means you can tailor the resulting sound to your exact requirements.
Let me explain. Figure 2 shows a block diagram of a typical VCF, and with luck, the more astute amongst you will note that this is not all that dissimilar from the block diagram of a typical VCO we printed last month. The only difference is that, because it's a processor, it has audio inputs as well as control inputs. It also has a summing amplifier to add a collection of incoming voltages: one of these is derived from a DC source, and it's this that sets the cutoff frequency shown in Figure 1. The other inputs are usually connected to controllers such as the low frequency oscillator, the keyboard and the envelope generators, a piece of operational versatility that allows for a variety of filter sweep effects.
You'll remember that last month we saw the effect a slowly rising and falling voltage has on a VCO. Apply that same voltage to the VCF, and you'll be able to sweep the cutoff point back and forth automatically in accordance with the level of the incoming voltage. What this means is that as the voltage increases, the cutoff point will do likewise, allowing more harmonics through. Similarly, the cutoff frequency will fall 'in sync' with the drop in voltage, and the sound will become gradually more mellow as the harmonics are attenuated.
It's important that the summing amplifier adds these voltages together to form one composite voltage, so the DC source acts as the lower extreme of a sweep while the modulation level determines the upper extreme. There is, unfortunately, no easy way of explaining this phenomenon in words, so practical experimentation is really the name of the game.
Just in case you're stuck for some ideas on how to go about doing this, here are a couple of guidelines. For starters, set the cutoff frequency control to about two-thirds and all the other controls to minimum, and hold down a note. Move the cutoff frequency control to its maximum position and hear the effect of the sound getting brighter, then do the reverse and note the result of that as well. You should be hearing a characteristic wah effect - if you aren't, you're doing something wrong.
Now reset the cutoff frequency to about two-thirds and increase the modulation level of your synth's low frequency oscillator. As the control is advanced, that wah effect should begin to occur automatically. And if you try moving the cutoff frequency control upwards, the automatic wah will get brighter. This is-because you're adding voltage into the summing amplifier, which results in the combined total of DC and LFO voltages increasing. If you try moving the cutoff frequency control downwards, the bright wah will become a duller 'woo-woo' sound, something not entirely unconnected with the fact that you're now subtracting the DC voltage and thereby lowering the voltage total. Note though that you haven't actually altered the modulation level at all in either of these examples, yet there's still quite a bit of sonic variation to be had from them.
All the modulation level allows you to do is set the width of the sweep, whilst the cutoff frequency control allows you to set the region in which that sweep is going to occur. The rule of thumb is that if you want an extremely wide sweep, you should set the cutoff control fairly low and modulation level high, while if only a subtle sweep will suffice, set the cutoff point to the region you want the sweep to take place over and keep the modulation level quite low.
There is one complication, however. Many LFO sweeps go negative for half their cycle (Figure 3), though this shouldn't cause you too many problems so long as you remember the summing amplifier principle. For example, if you set the cutoff frequency to halfway so that it gives about 2.5 volts, and then add a fairly extreme sweep from a bipolar LFO (that is, one whose sweep goes positive and negative in the course of one cycle) the positive voltage will be added to the DC voltage and the cutoff point will go higher as a result. As the voltage falls, the cutoff point will do likewise and the sound will become more muted. Before you know it, the voltage from the LFO will have arrived at 0 volts (halfway through its cycle), so that if you add the two voltages together at this point you'll get only the voltage set by the cutoff frequency control. As the LFO swings into its negative voltage, you can add this to the DC voltage so that, effectively, the voltage drops (ie. 2.5V + -1V = 1.5V). So in this instance, the cutoff frequency control sets the point at which the wah effect rotates. If your LFO's output fits this description, beware of setting the cutoff point too low, or you may find you lose the sound completely.
The other control you'll probably find at the control input of your filter is one that allows you to route the keyboard voltage through, and this is actually more important than it sounds. To explain, thus far we've been moving the cutoff point up and down but the pitch of the VCO has remained static. Imagine what would happen if that situation was reversed, so that the cutoff point remained where it was set but the pitch was varied up and down. In fact, the probable outcome is as follows. As the VCO's pitch is increased by your playing higher up the keyboard, the upper harmonics are attenuated, and as that pitch is lowered by, say, playing a scale right the way down to the bottom of the keyboard, more harmonics are allowed to pass through, resulting in a sound that's brighter at the bottom than it is at the top. If you're still feeling a little bit on the lost side, have another look at Figure 1 and imagine the fundamental and its harmonics moving to the right of the page (upwards in pitch, in other words). It follows that if the cutoff point is left where it is, then it's only a matter of time before nearly all the harmonics are removed. And not surprisingly, the same applies to the reverse of that situation.
The answer to this problem is simply to route the keyboard's voltage to the filter control inputs, so that the cutoff point follows the pitch of the VCO. Still unconvinced? Try setting the cutoff point on your synth to between two-thirds and a half, then start playing the keyboard across its full width with the VCOs set at 8' or 4'. Note that the sound will become duller at the top of the keyboard with your synth's Keyboard Track control switched off or at minimum, but that the output will be evenly toned across the entire range of the keyboard with this set at 100%.
It's often desirable to mute the top end of the keyboard a little, so most manufacturers have seen fit to provide a variable tracking control of some description, but it must be pointed out that some synths have only a switch that selects On, Half or Full. Obviously not as versatile, but it'll serve much the same purpose.
There is in fact a further control that's associated with the VCF but is not usually voltage-controllable, though some modular synths do offer this facility. The artifice in question is the Resonance control, also known as Emphasis and 'Q' by some manufacturers. Its purpose in life is to sharpen the peak of the cutoff point so that harmonics in that are boosted (see Figure 4). What lies behind this innocent-looking control is a feedback loop in the filter that allows the sound to pass through the filter again, and it's this that creates the sharp peak around the cutoff point.
And the reason you get that characteristic dentist's drill whistle sound is quite simple. As you move the cutoff control, the harmonics at the cutoff point are boosted while those directly above and below are attenuated. Then as the cutoff point is moved, so the harmonic that was being boosted before is cut, and the harmonic that was being attenuated is boosted. With judicious adjustment of the Resonance control, you can hear each harmonic quite prominently as the cutoff frequency control is moved up and down.
My only comment on the use of high-resonance effects is the usual impassioned plea that you exercise care and taste. The sound has its place, but the Rick Wakemans of this world have done as much as they dare to flog it to death, so it can sound more than a little cliché. Not only that, your sound will lose much of its depth because the lower harmonics are cut slightly, and as you probably know as well as I do, once you've weakened your output, it can be mighty tricky trying to make yourself heard on stage, for instance.
Most synthesisers allow you to push the resonance up so far that the filter will begin to whistle and oscillate. What's happening is that the filter is starting to behave pretty much like a sine wave VCO, and can be used as such. In fact, it's also possible to feed your VCOs into the filter when it's oscillating to create various bell and other clangy sounds. A self-oscillating VCF can also be swept by the EG for various Simmons drum effects and other pseudo-percussive wonders, but rather than dwell on all these possibilities now, I'll leave you all sitting on the edge of your seats by saying that I'll be looking at these and other effects in more detail later on in the series.
There are, of course, many other kinds of sound filter in addition to the lowpass variety. Two more are illustrated in Figures 6(a) and (b), and they all operate in much the same way, however, and it's only the harmonic content of the filtered output that varies from type to type. A highpass filter for instance allows only upper harmonics to pass through unhindered, while a bandpass design will allow a blend of harmonics lying either side of the cutoff point to make their presence felt. Perhaps not surprisingly, use of either of these filters results in the creation of relatively thin, delicate sounds, due to the lack of bass information present.
Some of you may even be in the possession of filters with a variable response curve (usually switchable) which will allow you to select between a 12dB per octave roll-off and a 24dB one: Figure 5 shows the difference between the two. With a 24dB per octave (ie. 24 decibels of attenuation for every octave of frequency) setting, the degree of cut will be correspondingly more drastic, and it's for this reason that these filters are often referred to as being 'punchy'. A 12dB per octave filter, on the other hand, allows more of the harmonics above the cutoff point to pass through. And as a result, the sound is a bit more fizzy and trebly, though in reality, many experienced synth players and programmers maintain that the effects from either variation are much the same.
Feature by Steve Howell
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!
Please note: Our yearly hosting fees are due every March, so monetary donations are especially appreciated to help meet this cost. Thank you for your support!