Making More of your Ensoniq ESQ1
Article from Sound On Sound, March 1989
Want to add multi-mode filtering, additive synthesis, three extra envelopes, microtonal and fixed frequency tunings, and eight independent effects units to your Ensoniq ESQ1 or SQ80, without invalidating your warranty? Greg Truckell reveals how...
Awesome! That's what I thought when I first read about the Ensoniq ESQ1 synthesizer. I promptly sold my huge and cherished modular system, and joined the queue for an ESQ. What was so special about a synthesizer that it could knock the DX7 off the bestseller spot? It was, after all, really just an analogue-style synth with a few bells and whistles - wasn't it?
Well, yes and no. The ESQ1 had a programming architecture instantly familiar to any programmer who remembers knobs, which after the shocks of FM and digital parameter access could only point in the ESQ's favour. But there was more - literally more. The ESQ1 had more waveforms, more oscillators, more LFOs, and more modulation possibilities than any other keyboard in its price range. Of course, now we have the Kawai K1 and Roland LA synths with zillions of waveforms; but the ESQ still has points in its favour, even over these machines.
OK, so you know all this - you bought your ESQ ages ago, and you lived through the software upgrades, and maybe you even upgraded to the SQ80. The programming principles remain largely the same on both instruments, so I shall talk about the ESQ1, although everything I say applies equally to the Ensoniq SQ80 also. The ESQ1 really is a programmer's dream - so long as you know what filters, envelopes, and oscillators etc actually do, we can get down to business.
What follows can be regarded as a free upgrade to your ESQ. The new features I am about to describe include additive synthesis, multi-mode filtering, three extra envelopes, microtonal and fixed frequency tunings, and eight independent effects units, including chorus, reverb and echo. All it costs is a little time - and it won't invalidate your warranty!
The ESQ1's filter is just one of the advantages it has over the other synths mentioned so far in this article; the DX7 and K1 have no filter, and Roland's LA synths don't allow you to filter the PCM waveforms. On the ESQ, a waveform is a waveform, whether it's a sawtooth or a multi-sampled voice, and you are free to filter any waveform. Filter cutoff, resonance, keyboard scaling and two assignable modulators may seem generous, but whatever happened to multiple filter configurations? Lowpass may indeed be the most often employed sort of filter, peeling off the high harmonics, but wouldn't it have been nice to have at least the option of using the filter in a highpass configuration? Come to think of it, is one filter really enough, considering the complement of envelopes, LFOs, DCOs, and so on? So now you're starting to think that the ESQ really ought to have at least two multi-mode filters - well, don't panic. Many of the things you think the ESQ1 can't do become possible when you start to explore some of the powerful modulation possibilities.
Suppose you want to programme a sound with a perceived pitch which is equal to the fourth harmonic (that is two octaves above the fundamental frequency); brass sounds - real ones, that is - are often like this. What we need is two octaves of harmonics below the perceived pitch. With my huge, cherished, and sold modular synth system, I would simply have used a highpass filter to roll off some low harmonics, using keyboard tracking to keep the effect constant over the whole keyboard if required; I had plenty of spare filters to deal with high harmonics. Not so on the ESQ; there is only one filter. But think for a moment about what a filter actually does. It attenuates harmonics, yes? Now, Ensoniq have not only given us a wide palette of waveforms but, in the case of synthetic waveforms, they have even told us which harmonics they contain and at what relative amplitudes.
What we need for this sort of brass pad sound is a waveform containing harmonics 1-4 and a few more. This waveform will be used on OSC2 tuned two octaves below the main oscillator, OSC1. It will also be passed through a DCA, set to a lower level than the others. It would be handy if harmonics 1-4 could be of at least equal amplitude; if they were of decreasing amplitude (as in a sawtooth wave) then we might get a two-octave detuned effect, which may or may not sound OK as far as it goes, but we only want one perceived pitch.
The ESQ's PULSE2 waveform fits the bill, and is band-limited, so we can worry less about the effects of mixing high harmonics. SYNTH3 may be another possibility, though it lacks a fourth harmonic, which is in fact the perceived pitch of the whole patch; however, the other two DCOs would more than compensate.
Waveforms containing prime numbered harmonics, like SYNTH3 and PRIME, and to a lesser extent the other two SYNTH waveforms, come in very handy for filling in the gaps in a harmonic structure. Remember also that, if you set an oscillator an octave higher, you double the space between its harmonics relative to lower-tuned oscillators. For example, if your fundamental has a frequency of f Hz, and OSC1 is tuned two octaves above the fundamental, then OSC1's fundamental is 4f Hz, and OSC1's second harmonic is 8f Hz, and its third harmonic is 12f Hz, and so on - this is how gaps can appear in your harmonic spectrum if you use some octave detuning between oscillators. It's a useful tool - once you know it's there.
Have a quick glance at the patch parameters for 'SOSPAD'; the SAW and PULSE2 waveforms have been used on OSC1 and OSC3, with OSC3 tuned an octave below OSC1. The DCAs have been set so that Oscillators 2 and 3 have one-third and two-thirds the amplitude of OSC1, respectively. A little brain ache tells us that the loudest harmonic is the 4th (which is the perceived pitch). Next loudest is the 8th harmonic; then the 2nd, 6th, 10th, and 12th are equally loud, followed by the 16th and 20th, then a group containing the 14th, 18th and 22nd harmonics, and... (see Figure 1).
Of course, things aren't really that simple (or complicated). DCA2 is modulated by ENV2, so the first 11 harmonics will not have a constant amplitude throughout the sound. And DCA3 is modulated by LFO2 (which is modulated by velocity and its own envelope) using the NOISE waveform, so the first 11 even-numbered harmonics don't have a constant amplitude, but will 'shimmer'. The pitches of the oscillators are finely detuned and modulated to various degrees by ENV1, so the harmonics will be drifting slightly relative to one another, particularly at the start of the sound. Of course, then there's the filter, resonance, keyboard tracking...
All this talk of harmonics is starting to sound less like analogue synthesis and more like FM or additive - and we haven't really exploited much in the way of modulation yet. The points to be made from all this are (1) that you don't need three different sorts of filters to poke holes in your harmonics, and (2) that you don't need a Yamaha DX7 or Kawai K5 to programme with harmonic structure in mind. Think of the ESQ1 as a digital wave synthesizer, like it says on the front panel.
If you want a group of harmonics to be modulated, but the filter won't do the job, choose a waveform that closely matches the harmonics you want, and modulate the DCA it passes through.
Using this sort of technique, you can have all sorts of things happening to your harmonics without even changing the harmonic bandwidth of your sound. Don't forget, either, that you can use oscillator sync, with ENV modulation of OSC2 to produce swept harmonics, while leaving the filter for other things.
Four envelopes is plenty, right? I mean, some synths only have one or two. On the other hand, a Casio CZ has six and a DX7 has seven. Not to worry; the ESQ has seven as well. Didn't you know? No, it wasn't a software upgrade that you didn't hear about - the extra envelopes have always been there, on the LFO pages. Stop thinking ADSR for a while, and it will start to make sense. An envelope is used to modulate a signal over time. If all you need is a simple up-and-down affair then the L1/Delay/L2 ramp will do the job - select SAW or SQUARE LFO waveform, RESET on, and a slow value for FREQ, and you're away.
There are much more interesting possibilities, though. For a 'chiffy' type attack, so popular since the D50 arrived, modulate a waveform containing plenty of widely spaced high harmonics (eg. BELL, or any of the Additive waveforms) with an LFO using the NOISE waveform - you could also modulate that oscillator's DCA with the same LFO, and with a short, fairly percussive envelope (which could come from an LFO if the others are used elsewhere).
Already fairly well known is the trick of creating echo-type effects by setting an LFO to the falling sawtooth waveform (ie. the sawtooth waveform, using negative modulation to invert its polarity) and heavily modulating DCAs 1-3 with the LFO, each DCA's level being set to low or zero. When you get bored with that effect, why not try using a different LFO on DCAs 1-3? I have used this trick to create some lovely tinkly wind chimes, with the three LFOs set to different frequencies.
You can also create pretty crazy effects by letting just a couple of cycles of the LFO through the ramp, and using the result to modulate some part of the attack transient. A useful technique when trying to programme some specific modulation effect is to go to extremes first, then once you know what the modulator is doing, back it off until it fits into place. Permit me to elaborate...
Suppose that you want to imitate the attack transient of a fuzz guitar. One way to do this would be to rapidly modulate the pitch of one or more oscillators up and down an octave a few times during the attack period. The way to go about programming it would be to work on one oscillator at first, so turn off DCA2 and DCA3 (just like you turn the operators on and off on your DX7, or the partials on your D50), and set up a slowish frequency on the LFO, which should be set to modulate OSC1. Try using the SQUARE waveform on the LFO first, and set L1 to 63 and DELAY to 0, so that the modulation is regular, and in jumps. Next bring the MOD depth up to 49 on OSC1, when the pitch jumps an octave (or so) at a time. Now increase the LFO frequency until it is just rapid enough for you to be unable to distinguish individual cycles (try 63 or thereabouts), and then introduce L2=0 and a fast ramp of 60 or more (a high value for DELAY is a fast rate, not a large time). Have a bash at the 'SOSFUZ' patch, where LFO2 is doing the business - and get your foot off the monitor!
Right then, we've added three envelopes and another filter to your ESQ so far - what else can't it do? Hands up all the DX owners who said 'microtonality and fixed frequency across the keyboard'. Oh come on, there must be someone. Thank you that scruffy little man at the back - but you're wrong anyway. Setting both OSC MOD1 and MOD2 to KBD with values of - 63 will give you a quarter-tone keyboard. No big secret, no fanfare. You have to mess about with the Octave, Semi and Fine tuning to get your keyboard 'in tune' with something; in the example patch ('SOSMIC') I have set these parameters so that some of the C notes on the keyboard are, in fact, Cs.
Now that you've had a glance at the example, you will have noticed that not all OSC MODs are set to KBD -63. Well, I thought, why should I tie up all my OSC MODs to produce one tuning effect? That would give rise to a pretty lifeless din. The thinking cap was hardly on before I had the answer - good ol' Sync. OSC2 is forced to follow the pitch of OSC1, without having to tie up either modulator. Of course, this still won't give me any pitch modulation, but at least I have more scope for timbral tricks.
What about OSC3? Well, you could just set it to the same tuning as OSC1, perhaps finely detuned for some chorus. I thought I'd put two tricks in one patch, and give OSC3 a fixed pitch across the keyboard. Just set OSC MOD1 to KBD2, DEPTH -49, and MOD2 to KBD, DEPTH -63. Simple, really. You just mess about with the tuning to get the pitch you want.
I decided to use the KICK waveform, through a very short percussive envelope, to add a bit of 'pluck' to the sound. A final word in your shell-like before you try out SOSMIC - it won't sound any better than any other synth set to quarter-tones, until you think of a use for them that is (hands up the TX81Z owners who still haven't used them). Nevertheless, it's easier to ignor a facility you've got than it is to use one you don't have.
Well, these days it all seems to be done with built-in reverb. So why hasn't the ESQ1 got one? Well, to be honest, it's got eight. No, it wasn't another software upgrade that you missed - it's a feature that has always been there. There are a couple of synths kicking around these days that use this sort of envelope controlled pseudo-reverb - the K1, for instance. The trick is to arrest the final release stage of the envelope, slowing its decay to zero at some point, so that from there on the sound lingers a little, giving the impression of reverb. This would seem to be tricky on an ESQ1, as it has no 'after-envelope'; that is to say, there is only one parameter after the note is released which the user has control over, and that is T4 - the decay time to zero level.
You can achieve this simulated reverb effect in a couple of ways. One would be to use two envelopes to modulate your DCAs, only one of which has a sharpish T4 -just so long as the sharpish one isn't ENV4, which controls DCA4, and which therefore wouldn't let your reverb out.
Just as effective though, and more economical in terms of modulation sources and destinations, is to use an envelope with a high L3 and low (possibly even zero) T4 setting to modulate the filter. Remember that the reverberant signal should be quieter than, as well as less bright than, the main signal. Try out the 'SOS' patch, and see what I mean.
I nearly forgot the chorus unit. Probably one of the simpler tricks, this one. We all know how to finely detune the oscillators to get nice rich sounds. But have you considered using the spare LFOs and envelopes to modulate the pitch of individual oscillators, just a little, to create a moving chorus effect? Other tricks might be to use envelope modulation to create a patch which starts with a big, fat detuned sound and then moves into closer unison as the sound evolves (good for choirs, string sections, and brass sections). You could even set up your modulation envelope(s) in such a way that the DCOs come to unison after the attack transient, then drift slightly apart again later on.
The more adventurous among you - and that often means those of you who have less gear than many - might bear in mind that, as a multitimbral instrument, the Ensoniq ESQ1 is also capable of generating several different effects, on different timbres, simultaneously. Don't worry too much about running out of voices; using reverb, chorus and echo can easily fill out the mix to the extent that you need less voices anyway. This can also solve the problem about what to do with a multitimbral instrument which doesn't have separate outputs for individual timbres. You can't use external signal processing on discrete simultaneous timbres - but you can use internal trick effects.
More useful than any of the hints, tips and advice that I have offered so far is one simple thought: the ESQ1 really is one of the big boys as far as programming power is concerned. You might have bought your ESQ thinking something like 'Great, analogue synthesis is back', but there is a good deal more to the ESQ1 than there ever was to the old Juno-type analogue synths. So be prepared to invest some time to get the most out of your machine.
Better yet, invest in a computer-based visual editing system. My personal favourite visual editing system for the ESQ, or any other synth for that matter, is the Dr.T Caged Artist editor called ESQapade. Think of it: all the parameters on one screen (or at least that's how it should be; some editors I have come across have nearly as many screen pages as the ESQ itself!). Don't worry about artificial intelligence and random patch generation; these things are there to help intelligent experimentation, which the ESQ so richly rewards. So keep thinking 'awesome', and stick with it!
Greg Truckell c/o Stiletto Sound Systems, (Contact Details).
Gear in this article:
Feature by Greg Truckell
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!