Outboard signal processors have been sporting MIDI sockets on their rear panels for some time now, but just what can MIDI do for fx and how can you best use it? Vic Lennard patches into processing.
AS MIDI BECOMES A MORE INTEGRAL PART OF EFFECTS UNITS, THE POSSIBILITIES FOR COMPLEX USE ARE OPENING UP. USING MIDI YOU CAN CREATE THE SORT OF EFFECTS YOU'VE PREVIOUSLY ONLY ASSOCIATED WITH BIG STUDIOS AND BIG BUDGETS.
AS WE ENTER the 1990s, it would be true to say that MIDI has become a way of life for the vast majority of technology-conscious musicians. We plug in MIDI cables in much the same way as their audio counterparts, and go about the setting up of Omni mode, the various MIDI channels and filters with relative ease. OK, understanding exactly how pitchbend information is stored as MIDI data may not be everyone's cup of earl grey, but we can all use it in the course of recording our music.
And it's not only the keyboards, drum machines and sequencers that we're familiar with that boast MIDI ports. These days nearly all modern effects units have the mandatory MIDI sockets on the rear - and yet this aspect of them is rarely used to any great advantage. In order to better understand what use these signal-processing animals can make of MIDI, let's divide the matter up - four distinct areas should cover it nicely.
IF YOU HAVE used the MIDI on an effects unit, then it's likely that you've used it for remote changing of the unit's current program. By sending a MIDI program change command from a keyboard, sequencer or other MIDI device capable of transmitting one, you can select which patch - and hence which effect - you want to use for a particular song. The time taken in sending this MIDI message is a little over half a millisecond so you are unlikely to spot a delay.
The problems which usually occur here concern the different numbering systems used by different manufacturers. MIDI can handle 128 program changes but always numbers them from 0 to 127. Most effects devices number their patches from one upwards and most keyboards do likewise, so sending out program 5 from the keyboard on the correct MIDI channel will change the unit at the other end to patch 5 as well. Sequencers are slightly different; most of them give the number of the MIDI program, so selecting 0 will give patch number 1 and so on. This situation has been the cause of more than the odd problem before now. The situation is made worse by manufacturers numbering the buttons on their keyboards in blocks of eight or 16. Roland's D50 has eight banks of eight sounds, and can be set up so that selecting a patch sends out a program change to connected devices. Bank 1, sound 1 will select patch 1 (or MIDI program 0) - fine. But what does Bank 5 sound 3, correspond to? The answer is patch 35 (4 x 8 + 3). Now, if only I cut off my two thumbs I could count in eights - it would probably improve my keyboard playing...
Consider this situation. You have two effects units which you want to change at the same time, and are both on the same MIDI channel. The problem is that you want program 15 on one unit and 23 on the other. What do you do? Most effects units have a program change table built in which lets you set up a table to select a specific internal program on receiving a particular patch change number. For instance, in our case, we could assign the two programs on the different units to patch change number 4. When this is received, the units will select their respective programs.
Some devices actually offer both of these facilities, including Alesis' humble MIDIverb II. Own up, how many of you MIDIverb II users didn't already realise?
SOONER OR LATER you're going to have filled up all the user memory locations in your effects unit - what now? If you could copy the contents of the memory into a MIDI librarian of some sort then you could start to fill the unit's memory locations over again. And again, most of the newer units offer just this facility. By coding it as System Exclusive information, the data can be sent to any device which can record SysEx - either sequencers (most sequencers will store SysEx information) or specialist MIDI data recorders such as the Elka CR99 or the Alesis Datadisk.
Alas, no solution is ever perfect. Firstly, the chances are that you will only be able to save the entire memory and not individual patches - this is dependent on the device. This is not a helpful situation, especially if you want to organise various patches together for use in a song or set of songs. Unfortunately, one of the most powerful methods of organising synth patches - using a librarian software - can't be used to help us out as very few units are supported by such visual editors. Secondly, some effects units cannot actually dump their internal data without receiving a "request" command from a librarian. Finally, there's the problem of early MIDI-equipped effects units. Take Yamaha's SPX90 - it has SysEx dumping information listed at the back of the manual, but doesn't have a MIDI Out port. For those of you still unaware of its existence, there's a switch inside which changes the MIDI Thru port to a MIDI Out. Without going into the dreary details, suffice it to say that you can dump individual patches or the whole lot. It is possible that other devices without a MIDI Out may possess similar features, but it is rare for them to be documented. A great shame.
NOW WE COME to the more interesting stuff. The first area we looked at - that of remote patch changing - used to be the only way to change sound processes in real time, apart from actually twiddling the knobs. Using this approach, however, all the parameters in a program are altered, even if all you actually need to do is change a reverb decay time. This is because you're updating the whole program rather than altering the parameter in question. Not only is this a clumsy way to modify a program, it invariably gives audible glitches. The alternative approach is to be able to change values of selected parameters via MIDI.
"NEW MIDI EFFECTS UNITS ARE LIKELY TO HAVE MIDI CONTROL BECAUSE THE EQUIVALENT UNIT FROM ANOTHER MANUFACTURER HAS - FOR ONCE THE MARKET IS WORKING FOR THE MUSICIAN."
This can be achieved in one of two ways. The first is by using what are called MIDI controllers - the information MIDI uses to implement mod wheels, sustain pedals and so on. Some of these are essential to the general running of a keyboard (certainly the above two), but controllers such as the soft pedal and portamento switch/time are less frequently used and so offer themselves for other purposes. You can select functions on suitably-equipped effects units and assign MIDI controllers to them. The type of MIDI controller and the parameter should match up - there is little point using a MIDI switch to change delay time, for instance. So, in this way, soft pedal can be used to change the type of reverb from hall to reverse and portamento time can be used to alter the reverb delay time or the pre-delay. Yamaha's FX500 is a good example of a "co-operative" MIDI effects processor, as you can select two functions from the 28 on offer and use any of 83 MIDI controllers along with note velocity and aftertouch to modify them. Alesis' Quadraverb allows you to control up to eight functions simultaneously over MIDI, and includes the use of the pitchbend wheel. There are some devices on the market which let you "inject" MIDI controllers in case your controller keyboard lacks them. The Anatek Pocket Pedal allows you to add a MIDI switch and a MIDI pedal to your armoury, while JL Cooper's Fadermaster gives you eight faders, each of which can send out any MIDI info that you program for it. In fact, there is immense scope here and some of the more expensive units really go to town. Eventide's Ultra-Harmoniser H3000 (reviewed MT, April '90) allows you to set its delay time to the period of an incoming MIDI clock, which effectively ties up the delay time to the sequencer tempo. By using this, any tempo change produces a proportional change in delay times.
The second way in which some units use MIDI to change internal parameters is via System Exclusive messages. ART's Multiverb is one unit that offers this facility. However, control of functions is very difficult to achieve from a standard sequencer, and a dedicated computer editor is probably what's necessary. Another unit capable of utilising MIDI parameter control is the Lexicon LXP1. This unit has a separate controller (the MRC1) which outputs SysEx directly to the LXP1 but which practically doubles the cost of an otherwise budget reverb unit. There is a solution here, however, as generic software patch editors - such as John Hollis' MIDIman and Dr T's X-Or - can be used as a cost-effective alternative.
YOU'RE MIXING DOWN your latest MIDI masterpiece; you decide it would be a neat idea to put the entire stereo output through a 180° pan to create the effect of turning the image inside out and back again. Allowing for the time in which this effect must occur, you calculate that a 1.36Hz modulation rate is required only to discover that none of the available units can produce this particular value. The only alternative is to try and create this pan on a live mix manually. As usual in such circumstances, the panning effect is required towards the end of the song and if you muck it up, you're back to the top of the mix to begin again.
The perfect solution would be to record the front panel changes of the pan on a sequencer and then edit these to produce the precise effect. Is this technology with us? The answer is a qualified yes - it's here, but at a price which puts it well beyond the scope of budget effects units.
The Drawmer M500 Dynamics Processor is up to the job, but it doesn't come cheap. Nevertheless, put into master mode, any panel modifications are sent out in an intelligible form to any device capable of recording MIDI controllers. Changing the M500 over to slave mode and replaying the sequence will reproduce those front panel movements. It's certainly a powerful machine, handling compression, gating, fading, panning, de-essing, limiting and expansion - with all parameters for these effects being MIDI controllable. Eventide's H3000 mentioned previously is also a member of this group of elite effects units.
Without going into details as to how these units work, suffice it to say that within the MIDI spec there are MIDI controllers designed for coping with the data. MIDI controller No. 6 tells a device that data entry is about to take place, controllers 98-101 can select a particular parameter and controllers 96 and 97 are for data increment and data decrement respectively. Some controllers have no specific definitions and so can be used by a manufacturer for a purpose they see fit. These include controller Nos. 70-79 and 84-91, while controller Nos. 80-83 are so loosely defined (general-purpose controllers) as to be effectively included within this group.
Controlling your effects unit via MIDI opens up a wealth of possibilities in the MIDI studio - of almost any size. Even patch changing can be used to good effect if intelligently applied.
New MIDI effects units are likely to have MIDI control of parameters simply because the equivalent unit from some other manufacturer already has - for once the market is working for the musician, so make the most of this: check out the relevant section in the manual before buying.
Once you're used to incorporating this aspect of MIDI in your music you'll find the benefits are twofold. You will be able to obtain a level of sophistication in sound processing you thought would only come with a rack full of expensive outboard processors and you'll feel that your equipment grows with your music, rather than restricts it. It's a nice feeling.