EASING OFF ON THE THROTTLE THIS MONTH, THE EMPHASIS IS ON SLOWER RHYTHMS AND THE TREATMENT OF VOICES TO WORK WITHIN THEM.

LOOKING BACK AT the example rhythm patterns published over the past 15 months of On the Beat, it occurred to me that the one area which has so far received scant attention is that of programming for slower tempi. Rarely have the patterns featured in the series ever ventured much below the preferred 120bpm dance tempo - even those which were never intended as dancefloor rhythms. Of course, I can imagine many people asking what it is about slower tempi that they should demand substantially different programming techniques; is it not possible to simply reduce the tempo of the patterns which we have already looked at? In some cases it might be, but as anyone who has tried to slow down a medium to fast tempo rhythm will know, a pattern played at well below its intended tempo range tends to sound rather like what it is - a pattern played well below its intended tempo range.

More often than not, the effect is of the rhythm being held back. Instruments seem to drag and what were previously interesting little rhythmic figures become unwieldy, rumbling sub-patterns which take forever to resolve themselves, and cease to complement the main rhythm. Another problem is the interrelation of the parts associated with each instrument. Ask a drummer (you remember - the human kind) to move from a fast to a slow groove, and the chances are that as well as reducing the overall tempo, he'll modify the number of bass drum beats in relation to those played on the hi-hat. If he didn't, the pattern would suffer in the way I've already outlined.

This relationship between instruments can affect quite significantly our perception of overall tempo, and it is often used as a means of manipulating a particular rhythm track to give it a slower (or faster) feel without doing anything to change the actual tempo. It is something that DJs (good DJs) quickly have to familiarise themselves with. I've often seen a dancefloor empty when a dance track has been followed by a record which sounds just as energetic but which is anything from 10-20bpm slower.

Of course it might be assumed that being written for slower tempi, this month's examples are unlikely to be of much use in a dance setting. But unless you're into all-night, open-air acid events (in which case I hope the harassment from police and insects hasn't spoiled things too much for you), I'm glad to say that even the hippest clubs still seem to end the evening with a couple of slow numbers.

If, on the other hand, you've no interest in dance music, this month's patterns could still prove to be a pretty useful addition to your library. One of the interesting things about slower rhythm tracks is that they tend to be far less stylised than up-tempo patterns. Adding a couple of keyboards to each of these examples, I found I was able to move into widely-differing areas without needing to alter the structure of the patterns in any way. This is not to suggest that further experimentation isn't needed. I'm only recommending that you try each pattern within a variety of contexts before deciding where you're going to use it.

After the relative simplicity of the last couple of articles, we're back to a more demanding level of programming, but even so, you should find few real problems. I keep wondering whether I should increase the number of dynamic levels to four or five, but the prospect of designing different patterned diamonds which would still be legible after being subjected to the rigours of the reprinting process is enough to convince me that it isn't such a good idea. And anyway, anything which potentially reduces the room you have for tailoring a rhythm to your own needs is something I don't believe I should be encouraging at this stage. Suffice it to say, three dynamic levels represents a bare minimum and you should be attempting to improve things in this area.

Right, to work. We'll start with a fairly conventional groove in Pattern one which, apart from a triplet figure that crops up in the castanet/clave line at the end of the second bar, shouldn't present anyone with any difficulties. Despite (or probably because of) this, the pattern is exceptionally flexible and capable of being used in more or less any situation, providing you don't need the snare coming down on the third beat of the bar.

If you have the choice, more open, double-headed toms tend to sound better here. but whatever you use, make sure they are kept well below the level of the other instruments in the mix. I think I would also tend to opt for castanets in preference to claves in this pattern, but as they are absent from most machines, I can see most people being forced to use the latter - though you could try a high-pitched wood block or even a finger click if this is available.

A triplet-based pattern with a rather more open feel, this month's second example, has, nevertheless, got considerable "clout" thanks largely to the space which surrounds the snare beats in every other bar. Also confirming its role as the dominant instrument are the flams incorporated into the cadence at the end of bar eight - as a general rule of thumb, these should be programmed as close together as your machine will allow and then moved apart until you're happy with the overall result.

The usual tendency for the hi-hat and ride cymbal parts to mask each other is overcome in this pattern by interlocking the two lines so that the two never coincide. However, you might still encounter problems with a long-duration ride voice overlapping the closed hi-hat and you could find you have to alter the pitch of one of the instruments to improve definition. The ride bell line, incidentally, is designed to accent the conventional ride cymbal during alternate bars, but if this instrument isn't available you could try simply accenting the ride using a higher dynamic level.

Another fairly spacious pattern, example three has a rather more straightforward feel and could be used wherever you need the solid one-two of a bass drum at the start of each bar. Actually, though I say bass drum, as you can see, the two beats are made up using a low tom as well as the bass drum. This, in fact, is particularly effective and helps overcome the somewhat cliched feel which this kind of pattern often has. That said, the tom figure at the end of bars four and eight give it something of a "rocky" flavour which might not be appropriate in some contexts. In this case, these are fairly easily dispensed with, and it shouldn't be difficult to devise a couple of alternative cadences using other instruments, if these are required.

A more demanding pattern in every sense, example four has much more going on than any of the previous rhythms and features a couple of rather interesting programming ideas. The first of these is the set of three snare notes in bars 1, 2, 3, 5, 6 and 7 which, being given descending dynamic levels, are intended to simulate the effect of a quickly decaying drum beat. Despite only comprising three notes, the overall effect of this is quite convincing if you choose your levels carefully. It's certainly a technique I'd recommend you try even if you decide not to use this particular pattern.

The second idea is designed to complement the first and comprises an open hi-hat note which is left hanging during the decay period of the snare notes and then shut off by a closed note immediately after it's died away. Again, even if you don't use this pattern, this is a technique you might try incorporating into another rhythm. It works well providing you have an instrument with a fairly long duration. The two small snare beats which occur in the cadence at the end of bar eight, could, I suppose, be called grace notes in that they are intended to lead into the bass drum beats which immediately follow. Programming is simply a matter of raising (or lowering) their level until you just become aware of their existence - but no more.

Pattern five also has a lot going for it. It is kept ticking over nicely with a foot-closed hi-hat (an instrument you really should try to add to your sonic arsenal), and demands a long duration, ambient snare voice and a very deep, open tom sound in addition to the three conventional toms which surface regularly throughout the pattern. If you haven't got anything suitable for the fourth tom voice, you could try using an open-sounding bass drum with its pitch raised, or even a bass note from a synth. Chosen correctly, the note should be of long duration (extending as far as the next bass drum beat), and be capable of giving the pattern a dark, brooding feel which is extremely effective when set against instruments such as claves and triangle. The pattern also makes use of grace notes - this time on the second beat of bars 2, 4, 6 and 8 in the mid tom line.

Returning to a lighter, more open feel, Pattern six is an extremely effective groove with a very distinctive flavour, but which nevertheless provides plenty of space for whichever bass instrument will be used alongside it. Care should again be taken with the overlapping ride and hi-hat parts: though perfectly complementary, the two really do need to be kept as distinct as possible. The pattern also features a cowbell, not normally one of my preferred instruments. But it does work rather well providing it isn't too open and "ringy". A damped cowbell would be the ideal instrument, but again, these aren't exactly common on the average beatbox. The flam notes in bar 9 should be programmed in the same way as the snare in Pattern two - starting with them as close as possible and then moving them apart until the right effect is achieved.

Finally we come to two rhythms which both rely on the toms for their interest, but which have quite radically different feels. The first features a neat, descending tom figure which leads into the snare beats throughout the pattern and leaves room for only a single bass drum beat at the beginning of each bar. Though the small diamonds on beat 3 of the closed hi-hat line are the same as those I used to indicate grace notes in Patterns four and five, they are not, strictly speaking, used for the same purpose here. Rather, they are intended to shut off the open hi-hat notes which immediately precede them without actually sounding themselves. In practice, it might not be possible on some machines to reduce them to zero and still have them cut off the open voice, but as long as they're kept low enough in the context of the mix, this shouldn't be a problem.

The second of the two, Pattern eight, is an altogether faster groove at over 100bpm. I suppose it only just qualifies for inclusion this month. Having said that, it's the type of pattern which could prove very useful in those situations where something in between the usual slow and mid-paced rhythm is required - particularly as it is so adaptable. There's nothing out of the ordinary in terms of programming, but as the toms feature so prominently, it would be worth spending a little time choosing the right voices and adjusting the tuning interval to suit the rest of the rhythm track.

As you'll see, none of the patterns are so slow as to be useful in only specialised contexts. In fact, if you're not familiar with programming rhythms under the 100bpm mark, you'll probably be surprised how little they need to be reduced below this to sound quite slow indeed. You'll probably also discover that the usable tempo range is significantly less than it is for faster patterns. The reason for this is obvious when you think about it: a range of ±30bpm represents a far greater tempo variation of a pattern running at 70bpm than it does of one running at 120bpm.

In practical terms, it simply means that you have to be rather more accurate in your use of the tempo control.

In an effort to save space and cut down on repetition I have only included one grid for parts 'a' and 'c' of each pattern - as these are identical. But obviously, each example is intended to be programmed in the order a,b,c,d.

If you're using one of these patterns and you find yourself forced to modify the speed of an existing piece of music by too great a degree, then you'll have to start thinking in terms of slimming down certain instruments within the pattern so that they can run at a higher tempo. It's difficult to be more specific, but with a little experimentation what I'm talking about should become apparent. Whatever happens, you'll emerge with a far better understanding of how to program patterns for both fast and slow tempi, and hopefully come to appreciate the different techniques demanded by both.