A Journey Through Synthesizer History

Sound On Sound's first decade has been graced by synths that were revolutionary, synths that were fascinating, synths that were seminal, and synths that were simply loved. Julian Colbeck takes a personal look at 10 of the best from synthesizer history.

As far as I know, I was the first person to compile synth reviews with the aim of putting this category of musical instrument into some kind of historical perspective and, coincidentally, the Keyfax series of books for which I began the task of compilation is also celebrating its tenth anniversary, like SOS. Now, of course, everyone is at it, and a quick flip through recent editions of SOS and other UK publications reveals that 'old synths' are now being re‑reviewed almost as regularly as new ones.

When SOS publisher Ian Gilby asked me to trawl my fading memory for a likely list of candidates for this feature, my first thought was that I would be mad to stick my neck out like this and offer what will invariably be regarded as my personal Top 10. I want to emphasise that it is not — these are simply 10 instruments from my Top 100 — or possibly from my Top 50 — and no more.

So I have decided to avoid burdening you with another potted review of the Prophet 5, or of the Minimoog, or of the Juno 60, and certainly of the TB303. My selection was based upon a number of criteria, amongst them: whether I had new thoughts or information on an instrument that could be conveyed in my allotted word count; whether an instrument has some particular relevance to the UK; and whether or not an instrument has by now been retro‑reviewed to death.

Like all good authors, I will conclude this introduction with a good plug for the forthcoming Omnibus Edition of Keyfax from which much of the new information in this feature has been taken. That's where you will find my Hot 100!

ARP 2600

MONOPHONIC ANALOGUE SYNTHESIZER

Made by: ARP (USA) from 1971 to 1981

Designed by: Alan R Pearlman

Original price: £1350

Current price: £1200‑1500

Sound generation: analogue, subtractive synthesis

The only ARP instrument I ever owned was an Odyssey. It was the first synth I ever bought, and it took myself and both the salesman in Boosey & Hawkes in London a good half an hour to get a squeak out of it. ARP were a perverse company. Nothing ever came — or went — easy.

ARP produced the 2600 as a more readily available alternative to their first big hit, the large and complex modular ARP 2500. With the 2600, Pearlman had his beady eye on the education market, hence the handy completeness of its built‑in amplification system and the informative schematics on the front panel. The 2600 was designed to be understood and used. And it was and is.

"With the 2600 you never have to listen to the same sound twice if you don't want to," said Joe Zawinul in an ARP promo leaflet of the day. Some would argue that you cannot get the same sound twice even if you do want to, and that this is part of the 2600's charm and attraction.

The ARP 2600 has spawned an industry of debate, conjecture, folklore, and rivalry. Is a 'Blue Meanie' (one of the garish, garage‑assembled earliest models), better than a 'Grey Meanie?' Does yours have the old 'Moog' style filters, or the later filters hastily designed by ARP in response to an impending law suit from Moog? Or indeed versions of these filters that were amended and improved after ARP went out of business?

With a 2600 you're not just buying a piece of musical equipment but also a source of endless speculation.

Anatomy Of An ARP

The nuts and bolts of the instrument are packaged in an oblong case whose lid detaches to reveal a telephone exchange of sliders, switches, and patchbays. This is a 3‑oscillator analogue design with separate keyboard, complete with effects processing and amplification. There are three voltage‑controlled oscillators, a 4‑pole low‑pass resonant VCF, VCA, two envelope generators, ring modulator, sample & hold circuit, white or pink noise generator, a mic preamp input, unique(ly) confusing voltage processor controls, and a spring reverb, most of which can be manually cross‑patched (each of these modules has an input and an output, making it an excellent instrument on which to learn what each part of a subtractive synthesis instrument actually does) for an unending array of analogue sounds.

It's quite possible to stand in front of this hands‑on machine and manoeuvre sliders until you hear something wonderful (especially if, as The Underworld's Rick Smith pointed out on the Museum Of Synthesizer Technology video, you start feeding it with outside inputs, like short wave radio signals). Sample and hold effects (the classic bleep and blurp provider) can even be clocked externally. Though MIDI was not even a nightmare during the 2600's production, interfacing this instrument with the modern world is not generally a problem. Any MIDI‑to‑CV boxes should provide MIDI access.

Design Evolution

The 2600 changed quite a lot during its long production run, externally from the original distinctive blue panel and metal casing, to more subdued grey‑faced and black models, to the final models blessed with rather cheap looking orange markings. According to ex‑ARP personnel, the later models were considerably better made, their modified internal designs at least allowing them to be repaired, an option that the epoxy‑entombed circuits of the early models singularly fail to offer. But then there's that trade‑off with 'originality,' 'rarity,' and 'sound.'

One of the most important ARP‑implemented changes came in 1975, when a modification from fellow synth designer (then mere peripherals designer) Tom Oberheim was adopted on production models. The modification not only provides a form of duophony (one oscillator serving the low note, another the high), but also provides a delayed vibrato feature, and a choice of single/multiple triggering. In 1978, American engineer Tim Smith discovered a fault in the 2600's filter which can cause the instrument to sound, well, dull. Smith, presently half of ARP experts Weyer/Smith Labs in Billings, Montana, is quite happy to share the secret of his discovery with anyone who asks, and the result is a frequency response of better than 22kHz. Weyer/Smith offer a vast range of 2600 improvements and upgrades, including a 'fix' for the dreaded 2600 'thump', which occurs if you try to set fast attack and release times on the envelope generator.

Arping On... And On... And On

In its day, the 2600 was a valued tool in the hands of most top players, from Joe Zawinul (top lines from such tracks as 'Scarlett Woman' and 'Black Market') to Tony Banks, Edgar Winter, whose monstrous synth sounds on Frankenstein are generally credited to the 2600, to Stevie Wonder, who was built a modified version utilising braille. Today's users include the likes of Rick Smith, and Vince Clarke, who reckons this is the best synth for making stupid, funny noises with.

A few years ago, as prices were climbing up to the £2000 mark, the ARP 2600 was in danger of becoming the darling of the synth bore: sought after, bought, rabbited on about, but never actually played. Such a criminal waste seems to have been averted with the recent price slackening. This is a magnificent instrument to play. Long may people afford to and be allowed to do so.

Casio CZ101

POLYPHONIC DIGITAL SYNTHESIZER

Made by: Casio Electronics (Japan) from 1985‑1988

Designed by: Mark Fukuda & A. Iba

Original price: £395

Current price: £100

Sound generation: digital, PD synthesis

When Casio produced the CZ101, it was as unlikely as Dixons producing custom design vinyl turntables. Less surprising than if they moved into the vegetable business, sure, but a definite shock. Casio's business was consumer electronics with a musical slant. For such a company to build a synthesizer, never mind a damn good one, was both intriguing and thrilling.

Why did Casio do it?

Yamaha's DX7 had set the synth market on fire a couple of years earlier and Casio, whose every product is specifically and solely aimed at the mass market, must have figured the age of the mass market musical instrument was upon us. And for a year or two Casio were right; both Yamaha DX7 and Casio CZ101/1000 sold by the thousand. But these were bright flashes in a very small pan. Never mind the heat, Casio couldn't stand the size of the kitchen and within a matter of four years called in the pro synth raiding party and resumed hostilities with home keyboards.

The CZ101 entered the world as a bargain basement digital synth initially sold, as often as not, by a bemused Casio sales force, through the same channels as its home keyboards. But its origins were far loftier. The CZ101 had been hewn out of Casio's much vaunted Cosmo Synth concept, whereby all manner of instruments and devices were being extracted from a sort of space‑age wonder synth supposedly designed for synth king Tomita for his In Praise Of The Earth extravaganza (a classic piece of mid‑'80s money‑burning which had The Synthetic One dangling out of helicopters, conducting an electronic orchestra that was floating about on New York's Hudson River).

Viewed now, as then, the CZ101 looks a most unlikely candidate for success. It's small — tiny, almost — with reduced‑size keys, and has little in the way of pro synth aura. But it sounds great: raucous, piercing, and with plenty of scope for idle noodler and serious programmer alike to create new textures.

Going Through A Phase

The CZ101 is based on Phase Distortion, a proprietary brand of digital synthesis involving messing around with a waveform's phase angle (representing its basic shape) and applying multi‑stage envelope generators to all three of the fundamentals of sound: pitch, tone, and duration. PD — as it became known — rivalled, and to an extent has similarities with, Yamaha's FM synthesis. It's very different in one respect, though, in that on the CZ101 it is entirely possible just to fiddle about and get some very workable, musical results. Much of this is thanks to the healthy number of dedicated control buttons, and a small but informative display screen.

The three main sections are DCO, for the oscillators; DCW — 'W' for Wave — whose functions are tone‑bending and filtering; and DCA for the amplifiers. Each of these sections can be shaped by its own 8‑stage envelope generator.

On the DCO panel there are a number of basic waveforms to choose from: sawtooth (best for brass or string sounds); square (for woodwind or organs); pulse (for thinner, more nasal, reedier tones); Double sine (a slightly beefier version of pulse); Saw‑pulse (a real rasper). Any two of these can be combined, and in addition there are three resonance waves which are, as you might expect, facsimiles of those classic Moog burbling filter patches, full of harmonics and whistlings.

Having made a waveform selection you can re‑shape it using the rate and level‑based envelope generator, an operation that may seem tedious in print and in prospect, but which in practice is beautifully offered and pretty painless to carry out. Indeed, the CZ101 is a very thoughtful little instrument. There's an edit/compare feature, you can initialise settings, there's key follow, portamento, noise, ring modulation, and (quite a killer punch at the time), the instrument can be used multitimbrally, with four monophonic voices. In 1985, MIDI itself was new enough and multitimbrality was a rare and little‑understood notion. The growth of computer‑based sequencers gave an added boost to CZ101 sales, as this was (and possibly still is) one of the most cost‑effective multitimbral sound sources on the market.

The CZ101's quoted 8‑voice polyphony is not entirely fair, since almost all patches use the Line 1 + Line 2 routing (ie. an oscillator signal path of DCO, DCW, DCA) for thicker, more complex and interesting doubled sounds (and less polyphony). And if this is not enough, a feature called Tone Mix renders the instrument full‑fat monophonic. And although in‑depth, precision programming is a feat generally beyond most people, there's still a forest of literature, videos, organisations, software, and sound cartridges around for users to get help at any level they choose.

Cheap And Cheerful

CZ101s are notoriously reliable. Yes — reliable. Casio's service department rarely sees any, and when they do it is often simply a question of pressing the underside reset button, marked P, while the instrument is switched on, in order to reload the factory presets. The CZ101 can be powered from batteries and if you let these go flat the memories will fade.

So many CZ101s were made, and the price was so low to begin with, that this instrument remains a bargain buy, unlikely ever to filter up into the crazy price ranges of some 'retro' instruments. Still capable of turning in some great sounds, though, this is an instrument no idle collector should be without.

EMS VCS3

MONOPHONIC ANALOGUE SYNTHESIZER

Made by: EMS (UK) from 1968‑1994

Designed by: David Cockerell

Original price: £330

Current (new) price: £1500

Current second hand price: £500‑1000

Sound generation: analogue, subtractive

All synthesizers have an element of 'toys for the boys' about them, and the VCS3 is the quintessential gadget for spotty Herberts who were no good at sport and liked to spend their free time in the science lab playing around with bits of wire and Bunsen burners. The VCS3 looks like a nice piece of laboratory equipment, with its large control knobs and pin‑matrix patchbay encased in a real wood frame. Its penchant for spacey, whoop‑whoop‑blip‑swoosh type noises that are totally unusable in any musical setting — aside from rhythm or sound effects, at which it is totally brilliant — reinforces that image still further.

Seminal Sounds

Over its long and frequently illustrious life, the VCS3 found favour with most of early synthesis' shakers and movers: Eno during early Roxy Music days, Pete Townsend, who had the VCS3 trigger the organ sound on The Who's seminal 'sequencer synthesizer' track 'Won't Get Fooled Again', Todd Rundgren, and Jean‑Michel Jarre, who still has a phalanx of them in his armoury. Shunting forward a few decades, Julian Cope continues as a valued customer, and the Aphex Twin is a reported loyal fan.

Designed here in the UK, the VCS3 was instantly popular amongst the European avant garde for a number of reasons. It was affordable; it could supply an endless and unrepeatable selection of high quality, interesting electronic bleeps and blurps; and though not exactly portable, due to the L‑shaped design which makes it awkward to pick up or to box, it was comparatively small.

Noise Machine

With its independent three‑oscillator design (independent not only in range (frequency), and level, but also in waveforms and applications — VCO3 can be used as a sub audio LFO), plus white and 'variable coloured' noise generator, self‑oscillating low‑pass filter, ring modulator, envelope shaper, and even a dual spring‑line reverb unit, the VCS3's basic range of tools exudes quality in scope and performance. Except, it has to be said, in terms of oscillator stability. If all you want to produce are 'interesting' noises, then drifting oscillators can be part of the sound, but for an accurate pitch performance this problem makes a VCS3 a high‑risk animal.

Though the oscillator and filter panels should seem familiar to anyone with a degree of subtractive synthesis experience, the envelope shaper panel, featuring a pair of attack controls, decay/'off,' plus 'trapezoid' and 'signal' level controls, will probably fox most — including me. With its variety of triggering modes, and decay time settable from the patchboard, this simply has to be viewed as an area for experimentation by modern, ADSR‑inclined users.

Much of the pleasure in playing a VCS3 lies in the small flat part of the panel's patchboard pin matrix, which allows you to connect up each of the instrument's 'modules.' It's a bit like playing solitaire, and, for most, with equally unpredictable results. But it is the VCS3's ability to be cross‑patched in an almost limitless number of permutations and routings that accounts for its enduring mystery and appeal.

It'll probably come as something of a shock to discover a reverb unit, which offers delays of 25 milliseconds up to 2‑second reverbs, again voltage controllable via the patchboard, and stereo outputs which effectively let you set up and deliver two separate but simultaneous sounds. In the days before MIDI, the only way to expand your sonic horizons beyond the wire‑covered lump of metal you actually purchased was to drive (or be driven by) other audio sources. External signal inputs let you process signals from a mic or a guitar, and the control outputs let you trigger external sounds using the envelope shaper or third oscillator LFO.

The VCS3 was one of the first instruments to offer joystick control, governing independent parameters on its X and Y axes. Joystick or not, unless your live performances do not entirely depend upon defined Western musical pitches, the VCS3 cannot be recommended for live use. Tuning is inherently hairy, and the prospect of a hard‑working musician under stage lights sweating profusely into the patchboard as he leans over the controls is frightening. Frankly, the VCS3 is most at home at home, and to this end there's even a small internal speaker system.

The VCS3 was also made available in 'briefcase' form, as the Synthi A. The KS Sequencer keyboard can be fitted into the lid of a Synthi A (thus making it into what EMS sold as the Synthi‑AKS). This 2.5 octave touch keyboard (no movable keys) even throws a 256‑note sequencer into the pot. The many illustrious Synthi AKS users include Alan Parsons and Pink Floyd.

Currently EMS is assembling what it says will be one final batch of Synthi A machines — nothing to do with demand; there are simply no more ABS casings left after this batch. Similarly, the VCS3 is officially deleted, but only because EMS cannot find a local woodworker capable of producing the French‑polished afromosia hardwood case for a sensible price.

Korg M1

POLYPHONIC SYNTHESIZER

Made by: Korg (Japan) from 1988‑1994

Designed by: Mr Ikeuchi

Original price: £1099

Current price: £650

Sound generation: digital, AI synthesis

Korg's M1 has been the big success story of modern times. If this statement of fact surprises you in any way, you must either be brand new to the field — in which case, welcome! — or else you are leafing through the wrong magazine — in which case, please put this down immediately and pick up your copy of the Angling Times.

Korg had been trawling particularly bereft waters in the mid 1980s, and when the codenamed '731' was previewed at the 1988 Winter NAMM show in California, I think no‑one, not even Korg, would have predicted that what we were all looking at would become the world's best‑selling synth ever.

Central to this sample+synthesis instrument (or AI synthesis, as Korg calls it) is 4Mb worth of PCM‑sampled and synthesized waveforms. From here, sounds can be shaped using quite conventional analogue‑style editing parameters, the result being stored as the standard M1 currency of a Program. Up to eight Programs can be stored in a Combination, which are used either for stacked or zoned sounds, or as a way of storing complete multitimbral combinations with each Program on its own MIDI channel. Throw in a decent pair of DSP chips for good on‑board effects, and a sequencer, and there you have the M1.

Why has the M1 sold so well? "Sounds" is the most obvious answer. The M1's ROM waveforms, and stunning array of factory programs (not to mention endless third‑party ones), sounded far better than anyone else's at the time and continued to do so for a number of years. Its factory programs, replete with acoustic guitars that really sounded like acoustic guitars, haunting oboes, and fierce, chunky pianos (the piano on LondonBeat's 'I've Been Thinking About You', and a zillion other dance soundalikes), melting strings, and sonorous basses, took the world by storm in 1988. In the cold, critical, physical modelling light of today, there's a certain four‑squareness to M1 sounds. A lack of sophistication, perhaps. Yet their directness and sometimes quirky character can be extremely useful. Chances are an M1 will get you out of trouble. Even now.

Inside The M1

An M1 Program can use one or two 'oscillators' as its starting point, each choosing its own multisound, from full multi‑sampled pianos to samples of bells and pan flutes, plus snippets of sound like the famous 'Koto Trem' or 'Lore' (a sample of a jack‑in‑the‑box being wound up, donated by Steve Winwood's keyboard technician), to synth waveforms hoovered out of that earlier Korg notable, the DW8000. The range and quality of these fundamental building blocks give the M1 its character and Korg its success, and much of the credit for this must be given to the Korg voicing team in Japan and the USA, who assembled the list of PCM/DWGS multisounds and who also went on to programme the factory presets.

You can alter basic pitch, the relative volume of two oscillators (where applicable), you can go to work on a four‑stage pitch envelope generator, low‑pass filter comprising cutoff point, EG, velocity sensitivity, and keyboard tracking (but no resonance), a similarly‑endowed set of amplifier controls, modulation controls affecting pitch or tone, aftertouch and pitch/mod wheel. Surprising facts are that the filters have no resonance parameter, there's only low‑pass filtering, there's no interaction between oscillators (ie. no‑cross modulation or sync), and only limited LFO options. In fact, Korg laboured long and hard over the filter, as on the original design the filter had no EG intensity. Fortunately, along with the fix came not just just the intensity parameter but also the interesting positive or negative intensity option.

Your Flexible Friend

There's still plenty to do and plenty of tricks in the M1. You could try using the pitch envelope to give a single‑oscillator sound an added bit of front end 'interest' (ie. a fast slide up at the beginning of each note). Not only can you create some catchy sounds like this, you can do it without having to resort to a double‑oscillator sound, which halves the polyphony. Another good trick is enveloping out the body of a full‑sample multisound to use as a brand new 'front end' in a double‑oscillator Program. The tip of the koto multisound is a good example. Organ patches that use the rotary speaker effect can be dramatically improved by connecting a footpedal to Effect Control 1 (on the Global page) so allowing you to control the slow/fast 'Leslie' effect in real time. Similarly, you can adjust many of the M1's basic controls — filter cutoff, release, effects level, and so on — in real time by selecting one of these parameters on screen (eight of them are displayed on the main Program screen, and can be accessed simply by pushing the little round function button underneath), and waggling either the edit slider or a connected footpedal.

Effects are found at the end of the regular edit pages in Program mode, and range from reverbs to delays, to overdrives, to chorus, to rotary speaker, to EQ, and are clean, powerful, and fully editable. The only nightmarish aspect of the M1 is the routing of these effects, because effects and overall output routing are inextricably linked. Choices abound, whether it be series or parallel for the effects, or a separate output for each program. There's even a couple of erroneous Program edit pages that don't do anything, just to make things really confusing.

Only die‑hards use the M1 sequencer, 'limited' to eight tracks but offering quantising and editing plus one or two interesting options such as the ability to programme and store patterns — bite‑sized chunks of sequence — in a separate slice of sequence memory that you can insert into your sequence wherever and whenever.

A fact tiresome to both Korg and the rest of the industry is that an M1 could still outsell most other synthesizers, even here at the end of 1995. Production did in fact cease in 1994, the final batch gobbled up ferociously in days. A remarkable instrument.

Korg Wavestation

POLYPHONIC SYNTHESIZER

Made by: Korg (Japan) from 1990‑1994

Designed by: Ray Keller, John Bowen, Scott Peterson

Original price: £1499 (Wavestation EX keyboard)

Current prices: £600‑1000

Sound generation: advanced vector synthesis

Tenacity is a key requirement of a synth design and a synth designer. With few exceptions, the process from idea to product is a long and complex one, though few have been longer or more drawn‑out than the story of the Korg Wavestation.

The Wavestation can trace its roots back to the mid 1980s, when Sequential (Circuits) as they then were, released a highly innovative synth called the Prophet VS. A hybrid digital/analogue instrument, the VS's oscillators did not produce static waveforms but rather waves that moved, evolved, and changed from one shape into another. Sequential fell into disarray not long after the VS and the company was purchased by Yamaha in 1987. After a year of not doing a great deal (though Yamaha's SY22 was to nod quite distinctly in this direction) the Sequential team was binned by Yamaha, and Korg, scarcely believing its luck, offered them all jobs the very next day.

A brief guide to Wavestation history might be helpful at this point. First came the plain Wavestation keyboard. In 1991 came the Wavestation EX keyboard, featuring additional piano and drum samples. Also in 1991 the Wavestation A/D module was released, similar to the EX but in module form and with the addition of analogue inputs. In 1992 the Wavestation SR was released, a slimline module brimming with every Wavestation patch ever dreamt of.

At the heart of the synth is the old VS contention that the waveforms — not the processing of the waveforms, but the actual waveforms themselves — should be allowed to move, cut and paste into each other, evolve, and mutate. On the Wavestation this vectoring principle is joined by an ability to string endless numbers of waves together, sequentially (ha!) to form what long‑standing Sequential sound designer and current Korg R&D stalwart John Bowen (whose idea it was), called wavesequences. Bowen saw wavesequencing only in gentle, evolving settings, but Korg programmers in New York spotted its potential for hard driving, rhythmic patches. And so the full breadth of the Wavestation came into being.

Fun In The Waves

Standard sound currency on a Wavestation is a Performance. Performances are made up of Patches, which are themselves created by combining up to four oscillators each, complete with their own analogue‑style filtering, envelope shaping, and modulation parameters. Incredibly, you can use up to eight Patches in a single Performance, which might lead you to wonder how the Wavestation can cope with multitimbrality, seeing as it's only 32‑voice polyphonic. Indeed. This is not the instrument's strongest feature.

Vectoring comes in at the Patch stage: an oscillator mix can be drawn using the joystick, the envelopes used to fade them in and out, with sections looped along the way. You don't need to limit yourself to blending oscillators at the same pitch, either. Great fun can be had mixing harmonies, or dissonances, for madcap, arpeggio‑type effects. Vectoring can be as serious or as serendipitous as you like, with analytical approaches or blind, random waggling equally likely to throw up a result.

Wavesequencing, stretching the principle out over a time‑base, compiles a list of waves, plus a list of commands as to when each wave will start, how loud it will be, its tuning, and so on. Depending on the type of wave chosen, you can achieve dramatically different results. Wavesequences can be clocked over MIDI. You can even, using the good many 'vowel' sounds in waveform ROM, wavesequence up a form of speech. The story goes that senior Korg personnel were highly sceptical of this claim by programmers — until one wag programmed up a message telling them all to f*** off.

Although the Wavestation programming system is generally understood by most with a loose knowledge of subtractive synthesis and some spirit of adventure, the occasional booby trap does lie in wait, the most serious being the relationship between a Patch and a Performance. Although a Performance is the base currency as far as actual playing goes, if you alter a patch, all Performances that also use that Patch will be altered as well.

Arming yourself with the motto 'Store Everything Separately', Wavestation programming is nothing if not fun. On a very simple level you can call up a Performance and simply substitute new Patches. Solo, a feature that isolates a patch in this setting, is mega useful. It may take a little head‑scratching to figure out that a patch, when settled in a Performance, is now actually called a 'Part,' and as such comes complete with its own volume, delay start time, unison or polyphonic play mode, and MIDI controller responses (cue noise of hair tearing out). But this does allow for much flexibility and is not quite so difficult to understand in practice as it seems in print — with the exception of the Wavestation SR, which is impenetrable, thanks to its pin‑sized user interface.

The soul of the instrument is at patch level, where, having selected the starter waves from a pool of hundreds stored in ROM (365 on original Wavestation, 484 on EX), you can impose velocity‑dependent pitch envelope modulation, do some low‑pass filtering, use rate and level type envelope generators, and assign two LFOs.

The Wavestation also has a glittering array of built‑in effects, provided by two 55‑algorithm multi‑effects processors. Effects are applied at the Performance stage, and yes, you can interact with many an effects setting from the mod wheel, or aftertouch, or a foot controller.

Added Extras

On the original Wavestation there were — duh! — complaints: "Where's the piano sound? Where are the drum patches?" So spanking new piano samples and drum programs duly appeared on the EX model. In the process, new effects algorithms were added — a compressor/gate, and some vocoders. Later still, on the 2U A/D module, analogue inputs allowed you to process the signal from a mic, say, or direct from a guitar, through the instrument, with an incoming signal treated exactly as if it were one of the ROM starter waves. Very exciting for those who like life out on the audio edge.

The Wavestation has been, and continues to be, used by a variety of top professionals. Peter Gabriel had a unit permanently set up at Real World during the recording of Us, and most of the pads on Michael Jackson's Dangerous were Michael Boddicker Wavestation patches.

The version you're most likely to come across today is the Wavestation SR. Doing much other than playing the SR, even altering its MIDI channels, is a headache. But the SR is packed with first‑class Patches and Performances culled from the best that Korg's wizardish team of programmers, led by Jack Hotop, could muster. If you just want Wavestation sounds pure and (not at all!) simple, then the SR is the weapon to choose. For the serious programmer, the A/D is the only choice.

Moog Prodigy

ANALOGUE MONOPHONIC SYNTHESIZER

Made by: Moog Electronics (USA) from 1980‑1984

Designed by: Rich Walborn, Tony Marchese

Original price: £295

Current price: £225

Sound generation: subtractive, analogue synthesis.

You've heard of the Bootleg Beatles, and Bootleg gin? Welcome to the Bootleg Moog. This was what Moog personnel called the Prodigy, a design concocted on the quiet by Messrs Walborn and Marchese and sprung on the marketing team's desk weeks before the annual industry spotlight, the Winter NAMM show.

If the Prodigy came as a shock to Moog, it descended like a giant can of insecticide on the British‑made Wasp synth, which had been the big buzz all year in the UK. Cheap, cheerful, but without a moving keyboard, the Wasp opened up synthesis to a level of musicians who never dreamt they'd be able to own a synthesizer. Cheap, cheerful, with a moving keyboard, and a prestigious name tag, the Prodigy instantly snaffled all remaining 'beginning user' sales in the UK. Interestingly, in a market thus softened up by the Wasp, the Prodigy was a good deal more popular in Europe than it was in the USA.

The $500 Moog

Walborn and Marchese wanted to see if they could build a '$500 Moog.' And they did. Clean‑lined, the Prodigy is stripped of all extraneous features, offering two oscillators with a switchable choice of sawtooth, triangle, or pulse (Osc 1: narrow, Osc 2: square) waveforms that can be pitched within a two‑octave range. Osc 1 calibrates this 32' to 8' and Osc 2 16' to 4.' An interval knob can split the voices up to a fifth apart, and a sync switch smartly lashes them back together again for those searing, forced tones that remain de rigeur in the Jan Hammer school of lead synth playing. A small mixing panel offers independent control over the level of each oscillator, plus a master volume.

At the time Moog made much of its 'heated chip' technology — not a culinary term, but a method by which Moog had hoped to cure the perennial problem of drifting (out of tune) oscillators. Maybe, but the Prodigy made a negligible improvement in this regard, and prospective purchasers would be wise to leave an instrument on for a while to see how well it performs.

The 24dB/oct low‑pass filtering features standard cutoff frequency, and 'emphasis' (Moog's terminology for what everyone else calls 'resonance') controls, and a slightly limiting ADS envelope generator. The filter can track the keyboard fully, half on, or not, and is good for self‑oscillating whistly ghostly noises. Modulation is limited, with a square and sine wave‑only LFO capable of modulating either or both VCO or VCF. It's all admirably straight to the point.

Yet, remarkably, the basic sound of a Moog synthesizer is retained — not classically so, and not particularly flexibly so, but with portamento and a pair of mod and pitch wheels, you can crank out some nice squelchy bass lines and more. The oscillator hard sync sounds work particularly well.

The Prodigy is encased in wood, and its basic construction was of good quality, which means that units stand a chance of survival. Some of the panel hardware was of less sturdy stuff, so some knobs and switches might have been — or might need to be — replaced. No matter: current street prices would suggest that the Prodigy is still something of a bargain for those looking for a piece of history at a price they can live with today.

The Prodigy is not the best Moog ever made, but it was particularly successful in the UK, which is why I have chosen to look at this instrument here, rather than conducting yet another retrospective examination of the Minimoog.

Improving Your Prodigy

There is an upgrade available for the Prodigy. It's called the Interface Kit, and is now handled by the newly‑formed Moog Music Technology. The upgrade adds a greater degree of (like, some) interfaceability than the original design's lone audio output allowed at the time, and comprises Moog S‑Trig and CV connections. You can contact Moog Music Technology Inc at 210 W. Ave, Depew, NY, USA 14043. Tel: 716 684 1090. Fax: 716 684 1091.

Oberheim OB8

POLYPHONIC ANALOGUE SYNTHESIZER

Made by: Oberheim (USA) from 1983‑1985

Designed by: Tom Oberheim

Original price: £4418

Current price: £800

Sound generation: analogue, subtractive synthesis

The OB8 was the last of a great series of Oberheim 'OB' synths that spanned the OBX, OBXa, and OBSX, classic instruments which have graced numerous classic recordings, from Prince's '1999' to Van Halen's 'Jump'. In retrospect, Tom Oberheim felt the OB8 was "too perfect," lacking the grit of the earlier models, but most users are pleased to discover a relatively stable Oberheim synth. Unless you want to dice with death (and a large overdraft for an Oberheim 4‑ or 8‑voice), the OB8 is clearly the model to go for.

Hidden Depths

The OB8 has two VCOs per voice, each with sawtooth, pulse, and triangle waveforms. Oscillators can be tuned separately and then shackled together in hard sync for searing, hollow lead‑line sounds. The OB8 departs somewhat from the earlier OB design in its filter section, now offering a choice of 2‑pole or the more drastic 4‑pole filter slopes. An ADSR envelope generator is reserved for the filter. The filter is certainly precise, which probably accounts for Tom Oberheim's slight misgivings on the instrument in retrospect.

LFO modulation can be in triangle, square, positive or negative ramp, or sample and hold waveshapes, and used to modulate VCO frequency, or pulse width, the VCF cutoff frequency, or the VCA. Deeper into the programming pages, you'll find additional LFO functions for altering the LFO sweep into half steps, and 'un‑syncing' the LFO for out‑of‑phase effects.

The LFO can also effectively track the keyboard, and thus speed up as you play higher and higher. Portamento options of smooth or quantised travel, and even polyphonic portamento, are also found here. Oberheim obviously felt (correctly) that such in‑depth features would frighten off the prospective programmer, and so hid these advanced features in a 'Page 2' mode which simply activates a second set of programming parameters under the control of the regular panel knobs and switches, once you double‑press the 'Page 2'/chord button. The Page 2 concept caught on to such an extent that it rebounded somewhat on Oberheim, who subsequently had to offer a new front‑panel screen which signposted these secondary, hidden functions.

The OB8 came out on the cusp of MIDI, which was offered as a retrofit or as standard for software revision B onwards. MIDI was obviously a bit of a rush job, since even the official MIDI version can only communicate on channels 1‑9, with just program, program dump, and lever information (in addition to note information, of course) capable of being transmitted. A ray of sunshine here is that in split mode two sounds will be sent out on separate channels.

Patches come in either single or doubled/layered mode, with 120 of the former and 24 of the latter capable of being stored internally. A cassette interface was the original method of external patch storage, though the MIDI specification does provide for patch data to be dumped and loaded too.

LA Story

The OB8 represents a good blend of the player's and the programmer's instrument. The flipper‑type pitch and mod wheel isn't to everyone's liking, true, but the keyboard is firm and the sounds feel playable. And there's a highly groovy arpeggiator that can be clocked externally (via arpeggiator clock input jack) if need be. Producer Steve Levine was a confirmed OB8 user in the mid 1980s, reckoning the instrument to exemplify the archetypal 'LA' sound — glossy, expansive, and expensive. More recently, Italian super remix team The Rappino Brothers use OB8 as a mainstay purveyor of analogue pads and gate‑effected rhythm synth parts.

The OB8 was the hub of Oberheim's 'System', comprising OB8 keyboard, DMX drum machine, and DSX sequencer — quite the setup to have for a brief period until MIDI arrived and blew the need for such restrictive practices to smithereens. You don't buy Oberheim synths for precision accuracy and reliability, but the OB8 was by far the most stable and easiest to service of the OB range.

Osc Oscar

MONOPHONIC ANALOGUE/DIGITAL SYNTHESIZER

Made by: Oxford Synthesizer Company (UK) from 1983‑1986.

Designed by: Chris Huggett.

Original price: £699

Current price: £600+

Sound generation: analogue/digital, subtractive synthesis.

In order to sell, any product has to be at the right price in the right place at the right time. The British synthesizer industry is particularly adept at getting at least two of these wrong, but with the OSCar it might have actually got the lot!

The OSCar poked its rubbery nose into our lives in 1983 — just before MIDI came along, with a hefty old price tag, and only on sale in one (London) music store. Short of advertising it as radioactive, it's hard to see how OSC could have made things much worse.

And yet... roll of drums. And yet... hoist that flag. And yet... puff that chest out, the British OSCar became (and remains today) a highly‑prized item. A monosynth with unprecedented power and range. A digital synth with an analogue spirit (and, OK filters). A radical design in all departments.

Best Of Both Worlds

Physically, lumps of rubber serve both as section dividers and end pieces, giving the OSCar an endearingly industrial look. Operationally, the big thing about the OSCar is that you can tweak and store. In other words, it's modern enough to offer memories, but 'old' enough to know the value of dedicated panel hardware (even if the identically shaped and coloured knobs are so long that they obscure their job description, unless you are a giraffe). You can also produce custom waveshapes; you can set up, and link up, and externally clock, sequences; you have an arpeggiator; you have fat, Moog‑type filtering. And yet such attributes are offered in digital form. Only the filters are analogue.

Oscillators 1 and 2 can both be switched to triangle, sawtooth, square, variable pulse, or pulse width modulated waveforms, and can be set to a basic pitch range in one of five octaves — together or apart. The pair may be fine‑tuned a semitone up or down — together or apart.

Osc 2's detune knob can also be used to set up a perfect interval between itself and Osc 1, and then the master fine‑tune knob can be used as a transposer.

In effect, there are two 12dB/octave filters, which can be used in isolation or combined to produce more powerful 24dB/octave filtering. Filtering mode can be band‑pass, low‑pass, or high‑pass, with control over cutoff frequency and resonance. A parameter called 'separation' governs the filter cutoff frequencies, and in band‑pass mode gives you two separate resonances. In addition to the filter envelope generator (featuring standard ADSR parameters), which can be either normal or inverted, some excellent repeat effects, courtesy of the filter envelope triggering, are available. For instance, you could clock the filter envelope using the OSCar's internal clock, with speed governed by the tempo control (the same goes for the VCA envelope or, indeed, both at the same time). If blitzkrieg power is what you want, the filter can be overdriven — an additional function of the master volume control knob.

The LFO offers three basic waveforms — triangle, sawtooth, and square — and its six‑position control knob also houses three routing options for 'env', which utilises the filter envelope instead of an LFO waveform, 'kbd' for a variety of filter tracking effects, and 'R' which produces a random sample and hold pattern controllable by the LFO's rate control. You can produce delayed vibrato effects by using the intro control, and the LFO can be used to modulate either pitch or tone. The VCA has its own ADSR envelope generator, and both single or multiple triggering is possible, and programmable.

Adventurous types can construct custom waveforms by combining harmonics, additive synthesis‑style, building upon a fundamental pitch in a wide range of combined levels. You can assemble complex, atonal waveforms, using the slightly cumbersome method of using the actual keys on the keyboard as selectors. Such self‑created waveforms can be stored for use in later patch programming.

Although the OSCar was launched just before MIDI, the omission was swiftly rectified, and subsequent MIDI models went on to allow all of its 36 patch locations to be user programmed. A MIDI OSCar is fully 16‑channel, allowing patch change, wheels, sequencer, and arpeggiator info to be transmitted. A little‑known fact is that when linked to a regular polyphonic sound source the OSCar can trigger a MIDI‑connected instrument polyphonically although the OSCar itself remains duophonic. Most OSCars sold were MIDI ones, and only 50 or so MIDI kits were installed, which means that there must be around 250 non‑MIDI models somewhere. So they will stay, it seems, as the design of the instrument will not allow a MIDI kit to be installed as it can be on a regular analogue synth. As with any such idiosyncratic design, spares are going to be a problem. The OSCar uses a lithium battery to preserve programs, and if you disconnect the board in order to fit a replacement (which you cannot avoid doing and will have to do every few years) your programs will disappear into the ether. The moral here is to save programs either to cassette, or dump data over MIDI to a data recorder beforehand.

As mentioned above, the OSCar has both a sequencer and an arpeggiator. Sequencer information is stored event by event, using notes and rests. Practice perfects a left hand/right hand technique of note and timing input that can almost be real time, and seasoned users generally swear by the variety of inspirational rhythms and effects you seem to be able to extract from the instrument. There are 22 sequence locations, and you can also chain sequences (with specific patch changes). Since the OSCar is technically a duophonic synth, it's quite possible to accompany yourself 'live' over the top of a sequence. Editing is concise more than all‑encompassing, and the total storage capacity is 1500 events.

English Classic

The OSCar was used by a number of influential keyboard players, including Billy Currie, who featured the instrument on 'Love's Great Adventure', a track on The Collection, Ultravox's Greatest Hits album (the angular, harmonic‑laden linking sequence‑type part is copybook OSCar). Jean‑Michel Jarre used it extensively on 'Industrial Revolutions', and Stevie Wonder used it on his Skeletons album. Apparently Stevie has 120Mb of OSCar samples tucked away in his Synclavier library. In the 1990s ambient and techno world, it's more a question of who doesn't use an OSCar — names include 808 State and Orbital.

Roland D50

POLYPHONIC DIGITAL SYNTHESIZER

Made by: Roland (Japan) from 1987‑1989.

Designed by: Jimi Omoto

Original price: £1445

Current price: £550

Sound generation: digital, LA Synthesis

I don't recall my first impression of the D50, at a press demo in London, being an especially brilliant one. Against the backdrop of a contorted explanation of LA Synthesis, the new Roland synth sounded bright, and certainly breezy, but only the scale of the event — ultra plush — gave any clue to the power Roland obviously felt was lurking behind this new technology. The D50 went on to become the first big‑selling instrument to use what would, amongst us pundits, become known as sample + synthesis design — essentially a subtractive, analogue‑style processing of sampled waves stored in a bank of waveform ROM.

By the time the D50 hit the streets it had been re‑programmed by American programming wizard Eric Persing, and its range of multi‑textured presets, such as the whirring 'Digital Native Dance', 'Chiffer', 'Gritttar', or 'Fantasia' became an instant hit with the public, reviewers, everyone. I bought one immediately.

At the time I was doing an album with Alan Parsons, and I remember being incensed when the record I had just plastered D50 all over was delayed. Even then I knew that within six months everyone would have those textures and sounds.

D50 Deconstructed

The smallest unit of sound on a D50 is a Partial, a sound snippet which can be one of two things: a short PCM sample — chink, chiff, blip, or bleep — or a full synthesized waveform. Partials can be combined in one of seven structures — PCM+PCM, PCM+Synth waveform, and so on. Thus combined, a sound then calls itself a Tone. Finally, you can combine two Tones in either split or layered formation. The result of all this jiggery pokery is that most D50 presets are indeed dual voices which utilise four Partials. A quick bit of maths will then reveal the instrument's tally of 32 available Partials to produce, in fact, no more than 8‑voice polyphony at the end of the day. Even in 1987 this was a little on the stingy side. Fortunately, classic D50 sounds tend to be harmonically rich and complex and so lend themselves to being used for specific, generally polyphony‑light parts rather than for the sweeping gesture.

Thereafter, D50 programming sticks closely to traditional subtractive analogue synthesis lines. Sound modifiers include a TVF Time Variant Filter (which can only be used for the synthesized waveforms, and which you think of as a glorified VCF with cutoff frequency, resonance, and keyboard follow, along with a plethora of parameters governing modulation amounts). The 'time' part of the name alludes to the filter's dedicated multi‑stage, time and level‑type envelope generator. Equally specific envelope generators can be used for the TVA (amplifier), and these, along with many LFO modulation controls, can be used by all the Partials. Each Tone can also make use of its own programmable chorus and EQ settings. Reverb is applied globally to the complete final patch, from 32 preset types, including halls, chapels, gates, slapback echoes, and so on. The quoted 24‑bit reverb exudes quality even if reverb tends to be the first thing to be ditched in a recording situation — at which point the D50's 8‑bit samples tend to be exposed in all their grainy glory.

The attack‑portion PCM samples comprise 47 one‑shot, percussive samples, 27 looped samples, and 24 re‑worked loops of other samples, amongst which you will find chiffy flute attacks, flailing drums, piano front ends, guitar chinks, and more. The synth waveform side of the instrument is designed to play the part of a sound's 'body', something that can also change over time using the extensive envelope generators. Though you may not ascend the full heights of detailed programming, you can achieve a quite substantial level of customisation simply by altering the balance between two Tones, using the joystick.

Aside from using keyboard dynamics to control volume, tone, and modulation, there is aftertouch — helpfully, slider controlled — portamento, and a neat feature called chase play which delays the start of one sound within a dual or split pair of Tones. Though the D50 keyboard is capable of great bouts of player‑initiated expression, it is also one of the most notoriously unreliable actual keyboards Roland ever made. The problem is generally one of notes either failing to trigger at all, or jumping straight to full 127 MIDI volume. Most often this is down to dirty key contacts, which, if you feel reasonably brave, you can clean yourself by taking out each key in turn until you get to the offending note and gently rubbing said contact with isopropyl alcohol. A tip: after you have done this, even more gently rub the contact with a piece of paper. No, not sand paper you clot. Just paper. For some reason, a contact that is too clean, works (or rather doesn't) similarly to if it were too dirty, and the paper rubbing action then dulls down the contact to perfection.

The MIDI side of the D50 is well defined and easily controlled, though notoriously a bit sluggish in response. Splits or combinations can be sent out on different channels, and mono mode is available (good for guitarists) but multitimbrality is limited to two channels. Painstaking programming, including mapping two halves of a split Tone to separate areas of the keyboard, can squeeze a little extra zoning out of a standard instrument but if you want actual multitimbrality you'd better buy yourself the Musitronics MEX Multimode Expansion board, which not only multitimbralises the D50 but also doubles the internal sound storage capacity. The polyphony remains as it was, though.

Many ROM and RAM cards have been made over the years (there are some good newish ones from Metra Sound), none of which are cheap. A more cost‑effective method is to store sounds on disk via SysEx. Not clear in the manual is the fact that you must press and keep holding down the data transfer button until you press Enter, when loading or dumping sounds.

A Sting In The Tale

The D50 might always be associated with chiffs and puffs and/or bell‑like sounds, but in truth it is capable of much more. It can produce a quite excellent organ (programmer Chris Macleod programmed a blinder for Rick Wakeman on his [and my] ABWstint, which resulted in the Hammond taking an early bath), while lush analogue strings, pads of innumerable hues, and moody electric pianos should be within the reach of even modest tinkerers. Not so good are acoustic sounds, notably not acoustic piano.

Hands up who hasn't used a D50? Right. But just for the record it is all over Prince's Lovesexy, Genesis were extensive users, and so was Sting around Dream Of The Blue Turtles, and on and on and on. It's a little too early for the D50 revival, and I'm ashamed to admit that I've sold mine. But I took care to hang onto my extensive library of D50 sound cards and SysEx disks. We'll meet again, that's for sure.

Yamaha DX7

DIGITAL POLYPHONIC SYNTHESIZER

Made by: Yamaha (Japan) from 1983 to 1987

Designed by: John Chowning (FM), Phil Nishimoto, Hans Yamada, Karl Hirano

Original price: £1549

Current price: £500

Sound generation: FM digital synthesis.

Hands up who hasn't played a DX7? Hands up who hasn't owned a DX7? Quite apart from the musical importance of this instrument, the DX7 opened manufacturers' eyes to the possibility that synthesizers could sell. From DX7 onwards, the synthesizer industry grew up.

The DX7's gestation can be traced back to the late 1960s when Stanford University's electronic music composition teacher John Chowning began experimenting with vibratos. Chowning discovered that he could produce musically complex, harmonically interesting results by modulating (using high‑speed vibratos) one sine wave with the output from another, and thus by 1967 the seeds of modern Frequency Modulation (FM) synthesis were sown.

Over the next few years, Chowning rationalised his work to the point where it had sales value for a musical instrument company. In a story reminiscent of Brian Epstein calling round record companies for The Beatles in the early '60s, Stanford was turned down by all the major American keyboard manufacturers, turning to Yamaha more or less as a last resort. Keen from the off, it still took Yamaha almost 10 years from when they took out a 10‑year license on FM to actually produce an instrument utilising this technology (the preset and hugely expensive GS1). But in 1983 all hell was let loose. The DX7 broke every rule in the book — sounds, playability, price. Yamaha simply couldn't make enough of them.

Operator Assistance

The DX7 is a 16‑voice polyphonic digital synthesizer, offering 32 internal memories plus a ROM/RAM cartridge slot. The keyboard is not weighted, but it responds to velocity and aftertouch. Further expression can be extracted from a Breath Controller (Yamaha's own invention), a mouthpiece/pacifier affair by which you can more convincingly simulate the breath‑to‑tone response of a wind instrument.

Chowning's FM theories manifest themselves on the DX7 as a series of 'operators' (which can be thought of as oscillators), that the instrument offers to the user in a number of different configurations or 'algorithms.' The operators are all sine waves, and can either be a carrier wave or a modulator wave, depending upon their position or relationship with each other in a particular algorithm. The DX7 can use six operators per voice.

The DX7 panel has plenty of dedicated controls, though switches are of the squishy 'membrane' type and the display screen is minuscule. Along the top of the panel runs a collection of algorithm diagrams, so you can see at a glance the type of sound you're likely to produce using each of these (effectively) oscillator configurations. An envelope generator, and a keyboard level scaling graphic are perched on the end.

DX7 programming is commonly, though not a little unfairly, perceived as impenetrable. Perhaps it was unwise of Yamaha to splash about words like operator and algorithm when oscillator or voice or shape might have been less intimidating. In any given program the novice programmer can simply switch operators on or off and begin to learn what role each performs within a sound, but things do get rather more involved when you consider that each operator can also specify a particular pitch, volume, envelope, and such, and so is almost a complete mini‑synth in its own right. In a flat, operator‑plus‑operator algorithm, the system can work as simply as drawbars on an organ, but once operators begin to interact, results become vastly more complex and unpredictable — which is why the system sounds so good, of course. Nonetheless, manoeuvres such as changing an envelope setting at the top of a stack of interacting operators can exert all manner of unexpected influences on operators further down, and it is this level of programming intensity that has thrown up the theory that the three essential ingredients of FM programming are trial, error, and luck.

It's not so much individual parameters that befuddle so much as the general level of interaction. Indeed Yamaha does retain many analogue‑style features and terminologies. The LFO, for instance, offers triangle, saw‑up/down, square, sine, and random waves, and can be set in terms of speed, delay, routing (pitch or volume), and amount. Although the envelope generators, which Yamaha bravely but wisely entombed in silicon rather than taking the 'flexible' software route, have their rate and level system emblazoned on the control panel, such multi‑stage envelopes are notoriously complex to set, especially without any help via movable graphics on the display screen (the tangible benefit of hardware envelopes by the way, is speed). If I may precis a section of Howard Massey's excellent book The Complete DX7, one rule of thumb is to remember that envelope generator control over a carrier will affect volume over time and envelope generator control over a modulator will similarly affect tone. There is also a separate four‑stage pitch envelope generator.

Many parameters that are sewn into the fabric of a program on more modern instruments are global parameters on the DX7. These include pitch‑bend range, mod wheel assignments, aftertouch response, glissando, poly/mono assign, and so on. These are all called Function parameters, and are accessed using the same buttons as the presets, or voice programming parameters. Said buttons thus have three separate purposes.

The DX7's superb presets — which must take a good deal of credit for sales — were programmed by two consultant programmers: Dave Bristow in the UK and Gary Leuenberger in the USA. Bristow and Leuenberger extracted every ounce of musicianship from the DX7, from the classic Fender Rhodes electric piano facsimile that has since entered synth folklore as 'DX piano' by all subsequent copyists, to the sonorous collection of fretted and fretless basses, to hand percussion, to bells, to marimbas, to ripping brass, to sound effects.

Lunchtime Legend

There can hardly have been a player in the 1980s who did not use the DX7 or one of its multifarious spin‑offs. There also grew up a whole industry of DX7 add‑ons and support products, including Grey Matter's E! expansion kit, which bolsters the patch tally to 320 while improving the MIDI spec to include local on/off, full 16‑channel access, and wide‑ranging MIDI filtering, and also adds some simple tone controls and the possibility of patch layering. A similarly user‑installable mod, DX Super Max from Group Centre, beefs up the program count to 256, offers patch layering and function programming, but also adds the delights of a superb arpeggiator. Many of these features were later to be found on the DX7II, a bold attempt to rectify the occasional DX7 foible, especially concerning MIDI (the DX7 came out in the same year as MIDI and so is understandably a bit limited — only being able to send out on MIDI channel 1 and not having a local off setting are the two biggest problems).

DX7s have not yet become collectable like Minimoogs or Odysseys, though that day will surely come in spite of the 160,000 or so units that were sold. Synth folklore has it that Dr John Chowning received no money for his work. Not true, if my conversation over lunch with him a few weeks ago is accurate. He may not have made the $20m+ that Stanford has over the years but he appears well satisfied with his deal. To its credit, Stanford University spent a good deal of the FM proceeds on a new and beautiful building for CCRMA, the Centre for Computer Research and Musical Acoustics (known as 'carma'), which Dr Chowning has spent his life developing and as the director of which he retired in the Spring of 1995.