When you've got both hands on the keyboard and you want to add real‑time expression, you need an extra limb, an abnormally long tongue, or some practical advice from Martin Walker.
Over the last few months, you can't have failed to notice the amount of interest in MIDI controllers that's surfaced in SOS, with Paul Farrer's excellent two‑part feature on making sampled instruments sound more real, and the contributions in the letters page and news items concerning alternative MIDI controllers. Why, then, do so few people actually ever get around to adding external controllers to their system? I suspect that this is partly because we're all a bit lazy, but also because many people probably don't realise that they might already have a socket on one or more pieces of equipment, which would allow them to attach a cheap volume pedal — providing an immediate external source of MIDI control without their having to spend any more than a few tens of pounds.
Of course, the easiest way to add more expression to your sequences is to employ aftertouch, if this is available (polyphonic or otherwise) but, despite being invaluable for adding general expression such as vibrato or preset pitch‑bends, aftertouch is rarely controllable enough — at least on most of the keyboards that I've used — to add the kind of effects needed to achieve the nuances of a string section or the subtlety (hmm...) of a guitarist. At one extreme the effect is off, and at the other it's at its preset maximum level; despite the variation in between, accurately returning to intermediate settings is extremely difficult.
After the keyboard, the next‑easiest step to more expressive music is to use any in‑built mechanical device — a modulation wheel, pitch‑bend wheel or lever, joystick, or even the ribbon controller provided on some keyboards (the PolyMoog and Roland JP8000, to name but two). However, the limitation of these is that unless you're playing a one‑handed keyboard part, you simply don't have a spare hand to do the controlling in real time. And, although many people use these devices to add more expression to previously recorded sequences on subsequent passes, the result is rarely likely to be as emotionally expressive as generating it during the actual performance would be, as you do when playing an acoustic instrument.
If both your hands are occupied, the logical way to add more expression is to use one or both feet with a pedal, or use your mouth with a breath controller. No doubt if you have access to someone who is well versed in the art of the one‑man band, your elbows and knees could be pressed into service as well. Of course, drummers have used all their limbs for creative purposes for many years, and top drummers are miracles of independent control (despite what other musicians may tell you). However, those of us who drive a car have already finely honed our footpedal technique, so this would seem to be the easiest type of control for most people to adapt to for MIDI purposes.
Putting Your Foot Down
Once you've decided to add an external MIDI controller, it's not always plain sailing. Although nearly all modern synths and modules (and indeed many multi‑effects units) respond to external controllers, you still have to find a way of getting data from your controller into the MIDI device. If you're lucky, your master keyboard (or alternative means of inputting note data, such as a MIDI guitar interface) will have a quarter‑inch jack socket on its back panel, labelled something like 'Expression' or 'Control' — if this is the case, you can just plug in one of the purpose‑designed expression pedals from any of the major manufacturers such as Korg, Roland, or Yamaha (see the 'Selecting a footpedal' box) and you're in business. If your main MIDI input device is some years old, it may have an expression input that's simply attached to the audio chain to control overall volume, in the manner of most home organs. While this is still expressive, no MIDI data will be generated, and you will have to record your performance onto audio tape or hard disk for posterity. You may find that you have a second halfway scenario, in that a plug‑in pedal alters MIDI parameters within the unit, so that you can achieve filter sweeps, volume swells and the like, but that this extra MIDI controller data is not re‑transmitted via the MIDI Out, so that once again the only way to preserve your performance is to record it. You can still often use external MIDI controllers with this type of device — by mapping the appropriate internal parameter to External MIDI control, and adjusting it via controller data at the MIDI In socket — but you'll still need a way to connect the pedal to generate this data.
What you need, to add external MIDI controllers easily, is an expression input, and a device that transmits this data via the main MIDI Out. To check out your own equipment, try looking for a MIDI Implementation Chart, found towards the back of all MIDI device manuals. Look under 'controllers', and you'll see two columns of data — Transmitted and Received. If either Controller 7 or 11 is ticked under the Transmitted column, and you have some sort of expression input somewhere on the back panel, then you're in business.
Wot, No Socket?
If your main keyboard has no expression jack socket, you'll have to look a little further for a solution. If another of your modules or effects units has one, you can use that, but then a slight complication arises — in order to achieve real‑time control of a MIDI device, you'll need to simultaneously send it the MIDI data output from both the keyboard and the pedal. The obvious solution is to use a MIDI merge unit, such as the Anatek Pocket Merge (available from SOS by mail order). If you have more than one MIDI input on your sequencer, you could also try routing the keyboard controller to one input, and the second device (with the pedal plugged in) to a second input. On a PC, for instance, many people will have more than one soundcard, and this immediately provides you with multiple MIDI inputs. This does mean, however, that MIDI controller and note data will be arriving separately, and if you want them to control a single MIDI device you'll have to use a multi‑client interface (see my feature on PC soundcard daughterboards in February 1997's SOS for more details on these). This will allow both streams of data to be combined, so that they both appear simultaneously at the input to the sequencer. In general, a hardware‑based merge unit may be more reliable than using the software approach, but there's still nothing to stop you recording note and controller data simultaneously onto separate MIDI channels if your sequencer allows you to do so.
The ribbon controller opens up many other possibilities, such as hammer‑ons and trills, that are impossible with any of the standard controllers.
If you find that none of your MIDI devices have expression control inputs, than you'll have to buy a device that can generate MIDI data from a pedal or other control device. One of the neatest ways of doing this is using the Anatek Pocket Pedal (again, available through SOS mail order). This little black box sits in line between your keyboard MIDI Out and sequencer MIDI In, allowing you to plug in one fully variable pedal, and one footswitch (on/off device). A set of DIL (Dual‑In‑Line) switches allows you to select the type of data output that the pedal and switch will generate. The options for the pedal are Volume, Pitch, Modulation, and Portamento Time, while for the switch you can choose Sustain, Sostenuto, Start/Stop, and Portamento On/Off. I tried one of these in line with my master keyboard, and it worked very well, effectively giving me a single MIDI Out with extra control inputs. The only slight disadvantage was that occasionally, at certain positions with the pedal left at a fixed point, the Pocket Pedal output oscillated between two adjacent values. Normally you wouldn't expect a 'stationary' pedal to generate additional data — only a moving one. Although this was too small a variation to give any audible glitches, the resulting continuous stream of controller data meant that I had to re‑patch the pedal away from the sequencer input during playback, as otherwise it kept overriding the previously recorded controller data, giving rise to all sorts of burbling effects.
Under Control
Once you have your controller in circuit and generating MIDI data correctly, it's time to patch it to control your synth parameters. The first and most basic use is to control volume levels (this is still vital for realistic strings and choirs), and virtually every synth allows this. In most cases, the external controller can be internally mapped to many different functions apart from this, such as filter frequency, modulation depth, pitch‑bend, or even feature multiple destinations, allowing much more expressive simultaneous control of volume, timbre, and modulation. However, if the external controller data emerges via the MIDI Out socket, it is often designed to appear as Expression controller 11 information, whatever the internal mapping. The easiest way to check just what is emerging in the way of controller data is to use a MIDI Monitor — either a software type (see the screenshot on page 80) or another MIDI device such as an Akai sampler. My S2800i has an extremely useful MIDI Receive Monitor that's ideal for examining data received from another device.
One thing that many people find confusing is the difference between controller 7 (volume) and controller 11 (expression). Although they both control loudness, and in many cases can be used synonymously, they may operate at different points in the system. For instance, in Yamaha's W5 and W7, controller 7 governs the volume after any insertion effect, whereas controller 11 is before the insertion effect; your choice here will depend on the effect you want to achieve. Roland say that controller 7 is used in GS sequences to set the overall volume of each of the parts (a sort of master volume), while controller 11 is used for real‑time changes (expression) between zero and the overall maximum setting provided by controller 7. I suspect that, in many cases, both controllers will act at the same point in the circuitry, so this is largely academic, but it's certainly easier to initially balance levels between different MIDI channels with one controller, leaving the full pedal range available for expression purposes using the other.
Looking at a selection of my own equipment to see what variations there might be, I found that my Yamaha DB50XG daughterboard responded to both 7 and 11, but that the channel volume (see the screenshot left) is the same as controller 7, so that altering either setting overrode the other, which might become confusing — the solution here is always to use controller 11 with a pedal. My Akai S2800i sampler only responded to controller 7 (which its MIDI Receive Monitor shows as 'Loudness'), but in this case the controller didn't override any other internal level control but acted entirely separately. My old faithful Korg M1 also only responds to controller 7, and, like the Akai, didn't duplicate any other internal value. Generally, since most recent devices are likely to respond to both controllers, it would seem to be easier to keep to 11 for expressive purposes if this option is available.
To optimise noise and digital resolution, try to ensure that most channels are used at peak levels somewhere near their maximum controller values. Incidentally, many modules (particularly the cheaper ones) have a default setting of 100 for the channel volume, because attempting to run all 16 channels with maximum settings of 127 could well result in distortion. If you are only running half a dozen channels or fewer, then whacking all of them up to maximum (127) value should result in a cleaner sound, since the background noise will be lower by comparison. And, unless you need to use it for purposes of balance, always set any master volume setting on your modules to maximum as well, using your mixer (rather than the module) to set final levels to achieve the cleanest signal.
Theoretically, as a pedal reduces the level, so the digital resolution will be compromised, at its worst just before the signal drops to silence. In practice, you're unlikely to notice any glaring reductions in audio quality, and many synths also incorporate some sort of noise gate at low levels. If this is the case, you may notice a tiny but sudden drop in volume at the very bottom end of pedal travel, rather than a smooth descent to silence. Judging by my own experiments, if you listen at low pedal positions, with the monitors turned up high, most synths will exhibit audible 'steps' in volume near the bottom of the range, due to the limited 128 steps allowed for 7‑bit controllers (0 to 127), but you're unlikely to notice this with other instruments playing in the context of a normal mix. Finally, don't forget that MIDI controllers can put out a huge amount of data — if you occasionally notice a glitch such as a note cutting off abruptly, or a sudden burst of high‑volume sound, perhaps you need to thin out your recorded data.
A Blow‑By‑Blow Account Of Breath Controllers
The breath controller is yet another way to add expression to a MIDI performance, and in the hands (or mouth?) of an expert can provide a great deal of subtle control as well as more extreme effects. Of course, synths specifically designed to respond to this sort of data, such as the Yamaha VL series, give excellent results, but breath controllers can be used with any synth to more accurately simulate a real‑world instrument. Breath controllers can control volume, filter frequency or amount of LFO; they incorporate a device known as a stress bridge. This is a small capsule that responds to pressure, which results in a movement of less than a millimetre; the bridge acts as a tiny potentiometer and produces a varying voltage that's converted into a MIDI data stream.
The problem lies in interfacing the breath controller. Unfortunately, the popular Yamaha VL70m has its breath controller input permanently connected to the sound‑producing circuitry; the breath control data does emerge via the MIDI Out socket so that you can record it into a sequencer, but you must re‑patch on playback or you'll get a MIDI loop, with the sequencer and the breath controller fighting to control the same parameters.
Built‑In Controllers
Everyone has their own opinion of how effective each type of built‑in controller is, but most manufacturers standardise on two thumbwheels — one with a centre detent for pitch‑bend, and a second for modulation depth. Often, one or both wheels will be sprung to return to their rest positions — for pitch‑bend this can make life a lot easier, but many people find it useful to leave modulation wheels in non‑zero positions for continuous effects. This two‑thumbwheel arrangement appeared on the original Minimoog, and most people find this combination comfortable, although the strength of any pitch‑bend spring is often a source of contention, and many people prefer a rubber‑coated thumbwheel for extra grip, particularly for live use when fingers can get hot and sweaty. On some six‑ and seven‑octave keyboards (76 and 88 keys), manufacturers have compromised by placing the wheels just above the left‑hand end of the keyboard in order to avoid an over‑long casing, but most people find these less easy to use. Today, most manufacturers still provide the standard two‑wheel arrangement, and many keyboard players have mastered the art of controlling both wheels simultaneously when adding expression.
There are several variations on this theme. Yamaha provide a third wheel on their VL1 and VL7 (as did Technics on their WSA1 — see below left), which normally controls a second source of modulation. Korg nearly always provide their dual‑axis joystick design, which is centre‑sprung in both axes and normally patched to provide pitch‑bend up and down when moved from side to side, and modulation depth (controller 1) when moved up. On my Korg M1 it generates controller 2 data when moved down. Controller 2 is now generally accepted as the Breath Controller standard, but on my M1 it's permanently routed to filter modulation. Although using a single 'stick' to simultaneously alter two parameters should give easier control, in practice there's sometimes some mechanical slack on either side of the central position, which feels slightly sloppy when you're trying to achieve subtle changes. Roland use their familiar 'bender' — another dual‑axis device, but this time the modulation (up/down) direction detects pressure in much the same way as aftertouch on a keyboard, and there's often no appreciable movement involved.
For some people, the ultimate built‑in controller is the ribbon, either mounted on its own, or incorporated into the famous 'log' on the Korg Prophecy, which is effectively a wide mod wheel (see above). The ribbon is an item affected by fashion — it has risen and fallen in popularity several times since its introduction by Moog as a separate device (much used by Keith Emerson during live performances back in the '70s). The new Roland JP8000 has already attracted much positive feedback about its ribbon controller, as has the Korg Trinity series, so perhaps the ribbon is making a comeback once again: certainly it's an ideal way to add real‑time expression, since you can use it exactly as you would the string of a guitar or violin — by directly recording any finger movement. In addition, the ribbon controller opens up many other possibilities, such as hammer‑ons and trills, that are impossible with any of the standard controllers.
Selecting A Footpedal
Nearly all keyboard manufacturers have at least one footpedal in their accessory range, but one possible problem should be mentioned at the outset. Although you may be tempted by low‑cost guitar volume pedals, some of these allow the internal potentiometer only 120° of travel over the pedal range, instead of its full 270° of available rotation. This will still give a range of volume control, but for MIDI purposes the output data will be restricted to about 0‑60 instead of the required 0‑127 values. In addition, guitar pedals are usually wired for 'through' use, with separate mono in and out sockets, whereas for MIDI use they normally need to be provided with a single stereo jack socket. If you have a pedal that provides the full rotation, and are into DIY, you'll need to rewire your pedal with the slider of the pot wired to the ring connection of a stereo socket, and the 'high' end wired to the tip contact. Most pedals incorporate a means of adjusting their stiffness of movement, and for synth purposes, this will normally need to be set so that the pedal can be left in a fixed position (for overall volume, for instance), whereas for rapid effects such as the classic wah‑wah, a sloppier setting will be easier to control, with the pedal flopping to 'full on' once you remove your foot.
Korg have the EXP2 Foot Controller at £69.99, and they sent me one to try out. It features both in, out, and expression sockets, and so could be used for general‑purpose swell pedal use as well as MIDI control. Rotation is over the full 270°, and it worked very nicely. The Roland EV5 pedal (shown left) is another one ideal for MIDI purposes. Yamaha market the FC7 pedal as well as the FC9 — in addition to the expression output, this features an in/out socket for in‑line volume control.
Sockets Included?
Many modern MIDI devices have jack sockets for the direct connection of continuous controllers such as foot pedals. The socket may be labelled variously Pedal, Performance, Foot Controller, or Expression, but they all tend to be the same thing — a quarter‑inch stereo jack socket for connecting a potentiometer‑style variable control, such as that provided by a foot pedal. In general, the keyboard versions of most synthesizers will have sockets for both pedal (continuous) and switch (momentary) controllers, although on the equivalent rackmounting module versions this is far less likely.
Korg say that most of their keyboards since the M1 have control sockets, including the current Trinity series, the Prophecy, the N264 and 364, X5 and X5D, and the Wavedrum. Roland products featuring a variable control socket include all their modern keyboards, most effects and even some modules. The JV1080 and 2080, XP50 and XP80 also store expression routing within their patch data, so that one sound might have it pre‑programmed to alter filter cutoff, while another could have it set up to affect resonance. Yamaha feature control sockets on most of their full‑size devices — the W5 and W7, QS300 and CS1X, for instance — although miniature keyboards such as the CBXK1 don't have them. Many Yamaha synths provide two sockets, one labelled Foot Controller for altering envelope, modulation or filter settings, and the other labelled Foot Volume, which can be set to generate either controller 7 or 11 data.
Most modern MIDI devices respond to both controller 7 (Volume) and controller 11 (Expression), but some earlier ones may not — those I've come across include the Yamaha EMT10 piano module (which remains totally oblivious to both), Akai samplers (which only respond to controller 7), and the Korg M1, which responds to (and transmits) only controller 7. Although the M1 has a pedal socket on the back panel that does generate controller 7 data via the standard MIDI Out socket, the values output when the pedal is swept from one end to the other only vary from 0‑99 to keep in line with its internal parameter ranges, rather than the standard 0‑127 range ideally needed for external use. No doubt there are other rogue devices like this lurking out there to snare the unwary — check your MIDI implementation charts.