You wouldn't attempt to play heavy metal on a harmonium (just to cover myself, I didn't say I wouldn't). The instrument would not readily lend itself to that particular musical form. Perhaps a Les Paul Junior through a Marshall stack might be more appropriate? The array of instruments available to the composer, songwriter and programmer is vast in its diversity, yet most have a fair idea of when to use a 909 instead of an 808, Marcato strings instead of a Juno 106.

The choice of instrument for a particular musical part determines to a large degree whether that part will work in context. Thus synth and expander manufacturers offer a wide canvas of timbres to choose from. Composers are individuals, each with his or her own unique view on how to put music together. The same applies to programmers, and here I include the people who write the software for sequencer packages.

Most sequencer users realise their musical ambitions on only one of the many sequencing platforms available. This is not surprising, given the cost, and lead-in time required to master, even one system. There is, however, a very good reason why you should take the time to investigate several options.

Sequencer manufacturers employ designers and programmers. These people tend to be mavericks, solving problems by creative lateral thinking. Since their products are largely a reflection of their idiosyncratic selves, it follows that some will more closely match your own working practices than others. You should, for example, certainly learn the basics of Cubase and Notator/Creator before shelling out for one or the other. There is also much to be said for stand-alone sequencers such as the Akai MPC60/ASQ10 or Roland MC50/500. While the latter lack some of the more outrageous features of their software brethren, they nonetheless score heavily for their ease of use, an attribute which arises largely from their bespoke control surfaces. Not for these machines the ergonomic straitjacket of the QWERTY keyboard. As with analogue synthesizers, where knobs and sliders that actually do something comprehensible have come to be revered, so some of the earlier models, with their array of dedicated buttons and data entry alpha dials, located with the musician or recording engineer in mind, are coming back into fashion. I use a Roland MC500 for sketching out ideas, and Cubase for high-level production work, and I find this combination offers pretty much all the scope I need.

An added bonus is that I can use the MC500's transport controls to control Cubase. Any tasty morsels recorded on the MC can be ported over to Cubase in real time, and worked up from there.

The scope of this article will be to outline, in broad terms the evolution of the sequencer, using as examples some of the devices I've used over the years. The important thing is to consider the merits or otherwise of each generic type, rather than drawing comparisons between individual models.

A POPULAR HISTORY OF SEQUENCERS

It all began with the arpeggiator. A pulse provided the timing and you played the notes. Any rhythm was possible as long as it was straight 16ths, and suddenly hi-NRG music was born. Rock-solid DIN-synced grooves hit the dancefloor, and it quickly became apparent that pop music would never be the same again. (Incidentally, DIN syncing is enjoying something of a comeback in certain quarters, as a rather retro solution to the problem of notes queuing up to be sent down a MIDI cable, to play ostensibly on the same beat). The Boss Dr Rhythm drum machine and Casio VL Tone (real-time sequence recording already!) were probably the first to round the corner to the home market, but neither could synchronise with anything else. While these units appeared on one or two novelty records (remember the drums on the awful 'Da Da Da Da'?), it took the collaboration of the Human League with Martin Rushent, using the first Roland Microcomposers (the MC2 and MC4) to announce the sequencer's arrival, doing all the tricky bits on the seminal Dare album. Most of the sounds came from the trusty Jupiter 8, and in many ways that album served as a showcase for Roland's then state of the art music technology.

Hot on the heels of this eloquent demonstration of what was possible (if you could afford the mortgage on some very expensive equipment; the Jupiter 8 cost around £4,000 when it was launched) came a couple of truly remarkable products, again from Roland, that didn't cost an arm and two legs. The TR606 Drumatix drum machine and TB303 Bassline sequencer offered a cheap but sophisticated complementary pair of DIN-synchronous units capable of producing a complete sequenced rhythm track, which opened the bedroom door on microcomposing.

ROLAND TB303 (1982)

Two oscillators (switchable), a filter with an audio fingerprint even more distinctive than a MiniMoog, and a monophonic step-time sequencer capable of (among other things) highly distinctive note bending. A number of these machines made their way into my recording studio.

It was noticeable that the style of programming varied surprisingly little between the various artists using the machine, all the more surprising because it represented a new musical direction (to contemporary ears, it would sound like the first hints of Acid House). Those early experimentalists were letting the instrument contribute to their musical expression, and as you are doubtless aware, the personality of the TB303 is particularly strong. Of course, it is arguable how much of this instrument's character depended on its sound and how much on its sequencer. My guess is that sound and sequencer take equal credit. (Incidentally, among those using 303s at that time were the band that have now evolved into The Shamen. And where are my West Coast Pop Art Experimental Band albums, guys?)

ROLAND MSQ700 (1984)

Hardly had the potential of using drum machines and sequencers together begun to sink (sync?) in, when along came MIDI (Musical Instrument Digital Interface, as if you didn't know). All of a sudden a ludicrous rumour began to circulate that machines made by different manufacturers could talk to each other. Remarkably, after a little teething trouble it became apparent that MIDI really did work, and the race was on to get your hands on a MIDI sequencer.

The first I had the pleasure of working with was the primitive, but easy to program, MSQ700. Capable of multitimbral recording in step time of up to eight polyphonic parts, this represented another quantum leap in compositional flexibility. In no time at all, anyone with a reasonably logical mind could master the basic principles: the chunky dedicated buttons had maybe two functions each, toggled by a shift key. The competition between programmers was to see how far the MSQ could be pushed beyond the limits of the manual ("for the time Signature you crave" being one of the more amusing instructions contained therein). I still don't know why the tape dump facility for storing programs only worked with the very worst cassette recorders. No matter, this machine could play a Juno 106 and DX7 simultaneously: to ask for more would surely have been ungrateful!

Of course, its idiosyncratic 'real-time' record/quantise functions led to much frustration among users, but its straightforward step-time recording enabled some fairly complex (if somewhat mechanised) compositions to be teased out of the machine.

Once the novelty had worn off, however, the search was on for a sequencer capable of recording reliably in real time, and quantising in a sensible manner.

ROLAND MC500 (1986)

The MC500 was the first of the next generation of hardware sequencers. Step-time recording was still possible, though less easily than on its predecessor. However, you could let that pass because this little beauty had true real-time recording, could handle up to 48 independently accessible tracks, had 'Microscope' editing, and, most importantly, sported a 3.5" disk drive for data storage and management. After a few hours with it, I had written a dozen or so pieces, of surprising complexity and diversity, if dubious musicality. Like the MSQ, its controls are chunky, taking their cue from tape recorder transport controls rather than the typewriter, and an experienced user seldom needs to look down at the modest and dimly lit LCD. This machine represents the last quantum leap in stand-alone sequencer design to date.

AKAI MPC60 (1987)

Akai's MPC60, combining a drum sampling workstation with sequencing facilities, is currently enjoying a renaissance among the brethren. I don't have one myself, but I can hazard a couple of guesses as to why this might be.

First off, like the MC500, it is easy to program. However, it has more advanced sequencing features than the 500, and can handle more tracks. It has SMPTE. It is also a drum sampler, and has one or two programming tricks specifically geared towards triggering its internal samples, such as a Roll feature. But the really significant selling points would seem to be:

1. Drum samples can be triggered from within the machine, thereby avoiding the potential timing problems that beset all non-integrated systems, introduced either by MIDI or by a sampler/drum machine's finite response time to incoming note data.

2. You can input rhythm parts as you would on a drum machine, via its generous pads.

3. It offers the highest degree of programmability available without incurring VDU eye-strain.

By this point there were a number of hardware sequencers on the market: Yamaha's QX series, Roland's MC marque, and Akai's ASQ/MPC family, to name but a few. Each had its pros and cons, but more significant are the similarities found throughout the genre. If you're looking for a hardware sequencer, avoid those without disk drives! The advantages of stand-alones over computer packages lie in their portability and reliability, together with the fact that they have control surfaces specifically designed for the job in hand. Of course they lack many of the features of...

COMPUTER-BASED SEQUENCERS

Obviously the most widely used type of sequencer today comes in the form of a computer program. Is there anyone out there who has not used Pro 24 at some stage? (Thinks — me!) With the most recent software revisions, the market leaders in the UK, namely C-Lab's Notator and Steinberg's Cubase, have tended to converge in terms of the facilities offered. They now differ largely in presentation. I rather like the scrolling of parts that are a feature of Cubase's graphics, but there is also much to recommend Notator's hands-on approach to its main page layout, with play parameters (transpose, velocity, etc.) directly accessible. (Though I note that Cubase v 3.0 now has a track play parameter list on the main Arrange window.)

Cubase, for example, offers graphic editing of notes and controller events on a variety of pages, such as Key Edit, Drum Edit and Score Edit. In addition to being able to process these in ridiculous detail, it offers ancillary functions such as instrument patch editing (including MIDI effects units), tape recorder transport functions (for the Fostex R8 and G16 only, so far), and other esoteric goodies like MIDI delay effects, and the Interactive Phrase Synthesizer, which allows the program to add variations around your theme. The MIDI Manager can represent any MIDI device, offering graphic control of not just Pan and Level, but also synthesizer editing. It can even learn how your synth works, and duplicate the controls. All in all, a range of features more extensive than most of you will ever know, far less actually use.

The ability of such programs to precisely define a groove of ever-increasing complexity and precision tempts the user to over-develop a small fragment of music, rather than broadly paint in a whole song. I feel this is one reason why popular music, especially dance music, has moved so far away from the traditional song format in recent years. Is it a case of not being able to see the music for the events?

As an example of how an advanced sequencing program can sometimes fall down on a simple task, consider this: it is not possible to drop in to erase notes on Cubase, as it is on almost any other sequencer. Since this method of thinning out a musical part is both intuitive and familiar to recording engineers, I cannot understand why Steinberg have overlooked it. Instead, when I want to delete notes from a bass line, or lose a dodgy chord, I have to open up an edit window and remove the offending event(s).

I contacted Steinberg for an explanation of why this should be so. Apparently, the difficulty in implementing drop-in erase on Cubase has its roots in the way the program calculates note positions. Cubase references all events from the song start (Absolute Time): the MC500 uses relative or 'Delta' time, referenced from the preceding event. The result of this is that while Cubase is able to offer Undo (last edit) as an option, the lowly MC can freely offer drop in erase/record at any time and place in the song, without the operator ever having to worry about defining locator points, etc. As far as I'm concerned, this little discourse only serves to emphasise the major thrust of this article. Sequencers are only human, after all!

Incidentally, while talking over these technical point with the man from Steinberg, I mentioned that I felt that over-developed sequencing biceps can muscle in on the music. He very sensibly quoted Pete Waterman, who, when asked "What is the most important piece of studio equipment?" answered "The first eight Beatles albums".

CRYSTAL GAZING

I'm waiting for someone to bring out a sequencer that combines the best of both worlds: powerful visual editing, but with a control surface that you can drive with your eyes closed. A machine that operates equally well on the events and the music!

STOP PRESS

Late news, just in! Drop in erase is coming to Cubase soon. I never did think much of that explanation involving Delta and Absolute times.