You learn a lot about people when you get involved with a computer. People are alternately intrigued and terrified by computers. Mention the word in connection with music and faces will either light up with fascination or drop in utter horror. Show them one, and they'll touch it and sniff it the way a primitive tribe might examine a white man who fell out of the jungle: what strange secret is it concealing? Can it talk? Is its skin slimy?

One followed me home once, and I decided to keep it. It was a PPG 340/380 Wave Computer from Mainz in Germany. What I've been through since has been a strange and stormy love affair.

Together, we're very formal and polite: I switch it on, fire it up with a shot of hexidecimals, kit blinks at me and greets me in an educated West German accent, and we immerse ourselves in a plan for World Domination in time for the next PRS distribution.

Apart, I'll curse it, sing its praises, laugh at the competition, sometimes wonder if I'm crazy even wasting my time with it. There's no doubt that it's changed my life: for better or worse is, well, academic. I'm an addict, and there's little I lean do about it.

When you consider that you can play tunes and sequence them with automatic percussion on a little Casio, that you could buy around a hundred Casios for the price of a Prophet 5 and several Prophets for the price of a Synclavier, you're only starting to recognise the enormous commitment involved in getting into computers and digital synthesis: just when you've arranged that second mortgage or signed your life away to some greedy record company in order to get your hands on that wet-dream machine, a brochure arrives through your letter-box describing a new Japanese wonder that does more for half the price (but won't be available over here for four months).

Guaranteed insomnia.

We're all pretty dependent on fragile machines we don't understand: for the man in the street it's that new washing machine, that video recorder, that microwave grill, all the gadgets he gets rammed down his throat by advertising and TV — until one of them claps out, and suddenly the bottom's fallen out of his world.

For a musician it's a music computer, which may not demand recording royalties or get drunk or go running after girl-computers, but can still leave him biting his nails and pulling his hair out. Which means, all in all, that if you don't want to end up a snivelling wreck you'd better think carefully before taking the plunge.

So what's available, and what does it do for how much? Music Computers fall into two main categories. First, microprocessor-controlled sequencers like the Roland MC4 or Oberheim DMX, designed to control external synthesisers; and second, complete systems which synthesise, sequence and perhaps sample live sounds.

It would take a thick volume at least to evaluate these systems down to the last detail, and by the time it was complete it would already be obsolete. The sequencers of the first category are easily accessible and relatively quick to get used to, so I'll try to assess sensibly the most important characteristics of each of the complete systems from the respective user's point of view.

Importers are starting to appreciate the importance of comprehensive demonstrations, so if you're seriously interested, I recommend you insist on spending some time familiarising yourself with a machine before pawning all your worldly goods.

I think it's important to describe some of the basic functions and objectives of micro-controlled musical instruments.

DIGITAL SYNTHESIS

This is a way of generating sounds based on the Fourier principle of fundamental waveforms and their harmonics (as opposed to the analogue system of oscillator and filter employed by the Minimoog, Prophet and most familiar manual synthesisers).

Basically, a digital synth is capable of memorising a number of tables of harmonics with different amplitudes, and calling them up in quick succession in order to generate a sound. The result is a much cleaner, glassier, more acoustic-sounding noise than an analogue synth, and the facility to imitate the character of complex real-life sounds.

SOUND SAMPLING

A digital synth can be equipped to accept an external sound input, say from a microphone or from a tape, and to analyse it and store it as a sound programme in its own memory. Feed in a second's worth of fuzz guitar, or a snare drum, or a dog barking, and it becomes possible to play that sound with a keyboard or sequence.

It's rather like a digital Mellotron: the difference is that you retain control over details of the tone, attack, decay and so on. The longer the duration of the sound you wish to sample the more memory the machine requires; so that a less powerful machine like the Emulator is limited to a few seconds worth of sampling, and it's necessary to take a start-and-stop-point within the sound and loop it for longer sustain; while the Fairlights and Synclaviers now provide a minute or more of sampling, so you can reproduce an entire sax solo in real time. A digital recorder uses the same principle.

SEQUENCING

A microprocessor can memorise a number of "events", or "steps", each consisting of a timing quantity, pitch, volume, sound group, and voice/channel assignment. By creating lists of event numbers and grouping them into individual sequences, and then looping sequences or running through them consecutively, a computer can perform a complex piece of music in any rhythm, tempo or key signature.

Sounds used in sequences can be percussive as well as purely musical, so drum/percussion patterns are fairly straightforward for most machines, especially if real drum sounds are samples and sequenced (viz. the Linn drum).

Alternatively, the micro can output voltage and gate signals to external synthesisers. For studio work, it's possible to lay down a sync code on multitrack tape to which the computer will always clock and synchronise.

This means that the computer's full capacity can be used for each overdub, and that any recorded line can be replaced or altered later, and that the computer can be run when mixing down from multitrack to stereo tape in order to save generations and available tracks by bypassing the multitrack altogether.

New music computers are appearing all the time, but for the moment the main contenders are the Australian Fairlight CMI, the widely-used American Synclavier II, the Emulator keyboard (also from the US) and the upcoming PPG Waveterm system from West Germany. I've crossed paths with all of these at one time or another, although I'm most familiar with the PPG, which has been developed from my own 340/380 system, of which only a dozen or so were made.

FAIRLIGHT CMI

The Fairlight is the most successful of the larger systems in this country, attracting the likes of Landscape, Trevor Horn, John Kongos, Robin Scott and others.

It consists of a Visual Display Unit, control unit with two floppy disc drives for memory storage, and a keyboard with a numerical keypad. At its most basic, it's an eight-voice polyphonic digital synthesiser with the facility for dividing its voices and assigning them to different parts of the touch-sensitive keyboard.

With its optional Music Composition Language, it's possible to create complex multi-voice sequences in real-or step-time, in any key or tempo. It has a sampling facility for anything from a piano chord to a bottle breaking and it displays its stored waveforms in an attractive form — an apparently three-dimensional graphic pattern that works a bit like a geological diagram.

Sounds can then be altered or re-arranged with the Fairlight's light-pen by touching the screen with the nib and simply re-tracing the desired waveform. The light-pen also makes general operation of the system a lot less tiring, because on most VDU's you have to run the cursor (a little blinking rectangle on the screen) down to the characters you're dealing with: on the Fairlight you just point to them with your light-pen.

The Fairlight is quite easy and pleasant to use — it's been well thought-out with the musician in mind. It speaks to you in quite a friendly way, using musical terms like "Bar" or "Tempo" rather than technical jargon, and points out mistakes without sniggering.

Most requirements are well catered for: it syncs to tape (by click-track or SMPTE code), has pedal connections and separate voice outputs with XLR connectors, and can be programmed with information relating to the keyboard touch-sensitivity and sliders for vibrato, glissando, attacks and decays.

So it's a nice keyboard to just fool around on, setting aside the programming aspect; and a good live instrument, with not too many buttons to hit by mistake. It could quite happily replace a whole rack of keyboards and effects, given that if you've got a favourite sound on your Prophet or OBX you can sample it into the Fairlight and include echo or repeats.

It seems to travel well: Geoff Downes of Yes took one on a 70-date tour with Yes and didn't encounter any problems, and John Kongos regularly packs his off to South Africa with him as excess baggage.

It has an integral audio amplifier, so all you need is a couple of Auratones and you can set it up anywhere. Servicing is fairly efficient, consisting mostly of replacing printed circuit boards that are bound to go down from time to time. Syco Systems (the UK distributors) will telex Australia for a replacement and it should be completed within days.

The most significant aspect of the Fairlight, and one which it's easy to overlook, is that to buy a CMI is a little like joining a club. You'll regularly be sent questionnaires to fill in, and the information they glean from each of their users is published and sent back to all owners.

Since they appeared in this country, Fairlights have been improved a great deal, mostly by software modifications brought about by suggestions from the users themselves. So you're never alone.

Perhaps this is what gives the Fairlight a marginal edge over the other systems, though it will only last until someone comes up with a machine at a more realistic price to the average keyboard player, which I can envisage happening within the next couple of years.

At the moment it's the domain of well-off and often detached and uninspired musicians and seems to encourage reclusive self-indulgence rather than earth-shattering innovation. Which is why a band like Depeche Mode, Soft Cell or Blancmange make far more exciting records than any Fairlight user in spite of their relatively primitive synthesiser systems.

SYNCLAVIER

This keyboard has been available in the US for some time where it's well-established. Over here it's just starting to take off: Heaven 17 have just bought one, and the Associates are taking one on the road. Francis Monkman swears by it.

New England Digital, its manufacturer, claim to have sold more of these than all other digital synth makers put together. They run a seminar outside Boston, Massachusetts once a year with lectures by people like Oscar Peterson and Chick Corea. Up until now they haven't sold many in the UK, though they've got rid of a few of the pretty brochures.

The reason is that, being as expensive as the Fairlight and nearly five times the price of a Prophet or Wave 2 (which does nearly as much), they've been fairly well eclipsed.

However, the Synclavier system has a potential which far outweighs anything else on the market, and now that the whole package is available, the Synclavier might win over a few buyers, provided they're prepared to part with the price of a perfectly decent London flat or a good album-and-a-half's worth of studio time.

The Synclavier system is based around its keyboard, which unlike the Fairlight, puts all its controls right upfront. In spite of this the keyboard is incredibly slim and light, weighing less than a Minimoog. It sits on its own legs and its one disc-drive clips underneath. The control unit is a black rack-mounting box that can be tucked well out of the way during performance.

The ergonomics of the keyboard are unusual: instead of a knob or slider for each parameter, you'll find a total of 108 red buttons. When depressed, a button will light up and the current value of the parameter it refers to will appear in a numerical window to the top left of the facia, expressed in terms of milliseconds, Hertz or decibels.

You then merely move one large knob next to the window in either direction, and the value shown will be increased or decreased. This is a neat system when you get used to it (the Moog Source has pinched the idea). There are two Morley-made pedals, one for volume and one programmable for various digital effects, and six foot-switches for all the usual hold/sustain/portamento functions, plus the arpeggiator and drop-in/out for the recorder (see below).

The keyboard can be touch-sensitive (for a few dollars more) although all vibrato, envelope, and brightness values can be programmed, and there is a Yamaha-type ribbon controller for pitchbending. You can store and recall an infinite number of sounds and sequences on floppy disc. So the Synclavier evidently possesses most of the performance controls your heart could desire.

So why the monstrous price-tag and why the surprisingly powerful 16-bit, 128 kilobyte microprocessor? Two reasons. First, the sound generation, which with its maximum of 128 voices and 96 harmonics puts the competition to shame. Secondly, the 16-track digital recorder concealed in the Synclavier's hand-rubbed African mahogany casing, which functions just like a multi-track recorder in most respects and can record up to 15,000 notes.

This is very easy to use — you just pick a track, punch up a sound, hit record and away you go, playing to a click-track if you like. Tracks can be edited, dropped in on with the footswitch, bounced together, transposed, or have their sounds edited while playing. You can loop sequences in various ways, though without a terminal to see what's going on, the mental gymnastics get a little awkward.

The only real problem with this system is that as the machine only has a mono output, all mixing has to be done internally, unless you transfer tracks individually onto multitrack tape — the Synclavier has a tape sync-pulse (which is no doubt different to that of a Linn, Fairlight and Roland systems).

I really wish manufacturers would put their heads together and devise a common code. The number of embarrassing hours I've spent on sessions trying vainly to hook up different machines with some producer cursing under his breath!

The new additions to the Synclavier package are the VDU terminal and digital analysis facility. These put it head and shoulders above the competition on a technical level, and accordingly rocket the price up into the megabuck range. The main functions of the VDU are: to give you a visual representation of sound timbres contained in the memory; to facilitate sequencing, so that music can be typed directly into the memory (which saves repetitive playing-in on the keyboard in some instances) and so that all notes can be displayed numerically; and to provide a manuscript-type display of the music, bars, clefs, time signatures et al. which can then be printed out in full for you to pop in the post to your publisher.

The digital sampling is at 50kHz and can sample up to 54 minutes of sound if the maximum memory is employed.

New England Digital claim that their machine will never be outdated. This may be the case, but I feel that their prices are a little prohibitive to mere mortals. They seem to be aiming at colleges, universities, film score writers and very rich American rockstars.

But it's a great machine to play around with, and you can rent one for a day from Keyboard Hire in London for merely the price of, well you know, a good used car.

EMULATOR

THE Emulator, by Emu Systems of California, is a really charming instrument. It looks very innocent, almost like the kind of kit organ featured on the front of Electronics monthlies, but underneath it's an interesting creative tool.

It performs the basic sound sampling functions of the larger computers without any of the fuss. You simply organise an input from a microphone or off tape, sample the sound and store it in one half of the memory (the Emulator only holds two sounds at a time).

You can tailor your sound slightly by picking a starting and stopping point, but basically if the sound is wrong, adjust it at source and sample it again. Once you've got two sounds you're happy with, you can play them polyphonically, together or separately, via a split keyboard. You can also sequence them in real time (around 2,000 notes).

When you're sick of them, dump them onto a disc in the integral discdrive, sequences and all, pluck out another disc and load up something else.

It's great for monophonic synth sounds you wish were polyphonic, or used as a studio effect for 10cc-type backing vocal "ooh's" and "ahh's" and strange repeats on individual syllables.

Unfortunately there's no way of synchronising the sequencer clock to other devices or even to tape, and I would have liked more than two sounds on hand, but maybe that would destroy the beauty of the instrument which lies in its simplicity.

The problem with computers is so often that spontaneity gives way to calculation and perfectionism: but the Emulator is very much hit or miss, and at least you can make quick decisions and retain your objectivity.

Does Brian Eno own one? I can just picture him wandering around the Kenyan jungle with an Emulator slung around his neck, playing plant noises back to them polyphonically.

PPG WAVETERM

This new PPG system is an extension of the Wave 2.2 keyboard. The 2.2 in itself is an excellent machine. It uses digital oscillators (two for each of its eight voices) to generate complex wave forms, but familiar analogue filters to treat them. This is a joy to keyboard players who miss the ease and speed of their old Minimoog or ARP when it comes to tailoring sounds.

It also gives the PPG a character quite unlike any other synth I've played: crystal clear, easy to slot in to the perspective of guitars and drums, and full of surprises.

It's not often that a keyboard manufacturer has the bottle to create something virtually unrelated to any of the competition. However, the brilliance of the machine is totally undermined by a disgraceful lack of service back-up and user co-operation on PPG's part, but I'll explain that in a minute.

The Wave 2.2 memorises 100 manually-programmed sounds, along with keyboard splits, oscillator detuning for effects, touch-sensitivity information, modulation values, and voice assignment (eight-voice monophonic, four-voice duophonic etc). In common with Prophets and Jupiter 8's, you can call up a program number and edit it simply by touching the relevant knob and adjusting the sound to taste.

However, all these values are displayed on the Wave's LCD crystal display (a bit like a digital watch-face) which is embedded in the facia, and values can be altered by typing a new quantity on the numerical keypad. The LCD display also gives access to the eight-track recorder, which functions quite like the Synclavier's but has one important luxury. Having recorded a line onto one track of the recorder, it's possible to play it through, adjusting as you go various aspects of tone, sustain and volume with the relevant knobs: and everything you do will be incorporated into the recorded line, and next time you play it back, you'll hear your expression in full.

You can build up eight tracks in this way giving you a very full sound — much fatter than a Prophet-10 or Oberheim DSX sequencer.

The Waveterm extension package consists of a terminal, an "event generator" and a sound-sampling unit. The terminal sensibly avoids using an entire typewriter unit by limiting itself to ten buttons whose functions vary according to the page you're addressing.

The name of each function appears on the screen directly above the button that controls it. You can use the terminal to examine all the characteristics of the Wave 2.2's sounds graphically, and it has a very ingenious trick of animating waveforms on the screen by calling them up in quick succession, creating the effect of an oscilloscope.

This is a useful way of viewing the structure of a changing sound. By using the sampling facility, natural sounds from mic or tape can be digitised and stored as sound programs, or adjusted beyond recognition. As soon as a sound is stored, it can be played polyphonically on the Wave 2.2's keyboard and embellished even further.

The event generator is the best sequencer I've ever used: it can sequence notes and bars in any configuration, move huge chunks of music around from bar to bar, handle all its own mixing, tempos and key signatures, zero in on one aspect of one note — or several aspects of a whole series — and make minute alterations.

Once you've overcome the problem of visualising your music as a series of events instead of fingers moving over keys or a fretboard, the possibilities are limitless. Having written a sequence, you can even ask the computer to write several variations, and sit back while it plays them to you!

But it's poor servicing which mars this otherwise superb machine. After-sales service and communication is absolutely essential on a machine of this complexity, and after owning a PPG for 18 months I have to say that the co-operation I've received has fallen well below standard. I would gladly have paid half as much again for my system, in retrospect, if only PPG had been a little more sensitive.

I've had problems with reliability on the road, and I've pointed out simple mistakes in the operating system which could easily have been rectified with updated software, but I've had hardly a peep out of the German factory. PPG's previous UK distributors Desert have experienced similar problems, but found the makers very reluctant to discuss them.

I hope Syco Systems, who are now taking over UK distribution, will have the necessary clout to sort this out, because it will prevent PPG from stealing the limelight away from the bigger manufacturers, which is a pity, because their machines deserve it. The Waveterm costs little more than half the Fanlight CMI and is in my opinion a more interesting instrument. I hope this throws a little light on a bunch of machines that although only in their infancy are already pretty mind-boggling. If you decide to fork out for one, I guarantee it will force you to radically re-evaluate your music, as well as wrecking your social life!

Computer programming is a pretty anti-social activity, but there again you'll be making music for people to slot in their Walkman and drift off to in private, so accept it as a symptom of this introverted era.

Future generations may be learning digital synthesis in primary schools and touchtyping ambient music into wrist-watch VDU's, but if you just want to plunder yet another Sixties classic and bring it "up-to-date", well that's okay too.

Music computers will be in every recording studio within a decade, so if you've got a friendly bank-manager, be the first on your block. But leave yourself enough to stock up with groceries.