Talking MIDI (Part 1)
Jay Chapman sets the scene for a regular series written to help you better understand the practicalities of life with the Musical Instrument Digital Interface.
It would be unreasonable to expect any magazine in the field of modern music not to describe, discuss and comment upon the Musical Instrument Digital Interface (MIDI) in terms of the technology involved and its ongoing effect upon the creation, playing and recording of music. SOUND ON SOUND certainly does not intend to shirk its responsibilities in this area so we commissioned consultant Jay Chapman to provide a series that will hopefully help readers to realise the full potential that MIDI has to offer them. Jay's first article presents a somewhat subjective and necessarily selective overview of what MIDI can now and might in the future offer the musician. Future articles will concentrate on the technical detail of how MIDI does what it does and how you can use MIDI to your advantage.
As you read this article, MIDI will have just celebrated the fourth birthday of its conception. Considering how ubiquitous MIDI is in late 1985 it may come as something of a surprise to realise that the MIDI 1.0 Specification in use today was finalised only two years ago (although some MIDI equipped instruments were already available by then) and was at that time a major departure for all the manufacturers involved.
The idea that synthesizers of various different functions and forms might usefully be connected together was not new, of course. In fact, the basic idea of connecting triggers, gates and control voltages in profusion comes directly from the first modular synthesizer systems where the voice program or 'patch' was to some extent the tangle of wires connecting the modules! In general, it soon became possible to connect devices from the same manufacturer so that a drum machine could drive a synthesizer's sequencer for example. Connecting devices from different manufacturers was considerably more difficult in a lot of cases and tended to force the musician into buying kit all of the same brand to avoid compatibility problems.
Whilst the manufacturers liked the idea of forcing brand allegiance on the poor musician, they lost as much as they gained since if the musician was already tied to some other brand he couldn't change! As new products were revealed old equipment might or might not be persuaded to work with it. Some standards did come into being such as the use of 1 volt per octave for keyboard CVs. Other manufacturers even managed to swallow their pride and use the Roland Sync 24 interface (and a trigger out, and a gate out, and...) which became a little more standard than some! Enthusiasts even produced trigger converters so that one synthesizer could talk to another. The mess grew and the musician got used to the idea of incompatibility and mourned the rapid obsolescence of his expensive equipment.
Here, it is worth pointing out that MIDI is not a standard in the sense that manufacturers are forced to use it by law! They are, in fact, coerced into using MIDI by simple commercial pressures - if their product doesn't have MIDI it won't sell very well: witness the availability, due to demand from the musicians, of MIDI retrofits for both Yamaha PF pianos and the mighty Fairlight CMI. Another important cost advantage is that manufacturers no longer have to sell expensive electro-mechanical controllers, such as velocity sensitive keyboards, with each sound source which obviously reduces the cost of a multi-synthesizer set-up and brings it within reach of a larger market. Both Roland and Yamaha have embraced this concept with their 'mother keyboards' (of which more later).
Whilst new technology, techniques and instruments must all have a considerable effect upon the musician's approach to his art, it is also necessary that he or she cultivate a new awareness each time a step forward is taken. Burying one's head in the sand simply makes you sneeze a lot! Some steps forward are larger and more far-reaching than others and yet are, at the same time, not immediately totally accessible to their intended user. MIDI is rather like an iceberg in the sense that it has so far revealed a mere 10% of its capabilities and uses. The 90% yet to be discovered will probably lie to a large extent in the province of 'in what areas will the existence of MIDI enable and promote development?' rather than 'what can MIDI actually do?' which is a question we can already answer in considerable detail.
Consider for a moment the fact that it is possible to play a synthesizer and even to programme it (to a greater or lesser extent) without a deep understanding of sound synthesis or how the synthesizer works. This is largely possible because the basic method of playing a synthesizer is similar to that employed in playing the keyboard instruments that came before it; as far as programming is concerned the empirical approach (ie. try moving the controls and see what happens!) usually yields some sort of results and lessons are slowly learned as experience is built up. MIDI, on the other hand, is a complex and yet largely intangible system-wide communications medium that interfaces with many aspects of sound production and recording technology. Its full potential is very poorly understood and musicians have little, if any, experience of similar systems.
It would seem that one can progress fairly quickly with MIDI by simply connecting MIDI Outs to MIDI Ins and thereby immediately produce some interesting results. Many synthesizer players with access to MIDI equipped instruments will have been delighted to use MIDI on the simple and practical level of making two or more synthesizers 'play along' with each other (Figure 1), thereby layering sounds. This mode of connection and use is not to be despised for its simplicity since it illustrates perfectly, but is only one example of, the central thrust of MIDI which is to permit one piece of equipment to control the actions of another (where such control makes sense!) whatever their manufacturers and internal details.
It should be obvious to anybody who hasn't been blindfolded over the past year that MIDI is not just restricted to the musical instruments themselves. In the context of a specialised 'musician oriented' network connecting performance controllers, sound sources and effects, MIDI simply acts as a standard communications medium which allows temporarily or permanently related pieces of equipment to function harmoniously together. An example of this is when a new effect setting is selected by an echo unit because a synthesizer has changed over to a new voice. This facility, which permits the musician far greater freedom and control of detail in performance, is quite new and its potential is as yet largely unrealised.
Once an instrument can be controlled remotely, be it by CV/Gate or MIDI, the development of dedicated sequencers comes on apace. Fuelled by the need of many a budding musician to play music beyond the capability of his technique, this technology, which is therefore an enabler for many, opens up new and stunning horizons for the virtuoso musician be he a player or more specifically a composer. The ability for one individual to control, in every detail and nuance, the playing of a multi-timbral orchestra in real-time (something which is beyond the technique of any individual) must revolutionise music and musicianship beyond our wildest dreams. It can also very easily trivialise and dehumanise a thoroughly human activity and yet by its simple regulatory effect, restructure and revise our thinking.
In fact, the major role that MIDI plays is that of explicitly introducing computer technology and, more specifically, the microcomputer into the world of music. This is not to say that computer science had failed to contribute to the musicians's lot before MIDI. Indeed therein lies a tale which we must explore if we are fully to understand the impact of MIDI and all that comes with it.
Sometimes the microcomputer is not directly on view or even mentioned in the instrument's specification (after all, just a few years ago a mention of computers might have frightened the sensitive and artistic musicians away). However, a substantial subset of the features of many current synthesizers result from computer simulations of oscillators, filters etc supported by more or less specialised hardware. The very core of sampling technology is, of course, based around microcomputers and serves to illustrate the versatility of the hardware and software combination. Computer Music has had its interesting moments in the research into computer-based composition, for example, but MIDI's role is different. MIDI currently enables or will enable the configuration and control of powerful individual elements and their combination into complex integrated musical systems.
It is the application of computer power coupled to a man-machine interface specifically researched and designed for musicians that is yet the goal of the (not too distant) future as far as the average musician is concerned. As the price of computer technology continues to drop so the computing power available to the average musician continues to rise. Current 8-bit microprocessor systems are hard pressed to cope with individual elements of complex music systems and the driving of a really excellent man-machine interface in real-time is beyond their capability. That is not to say that very reasonable attempts are not being made: have a look at the UMI-2B system reviewed last month or the Island Logic Music System which uses the latest graphic techniques combining windowing, icons and so on to make the system as easy to use as possible.
Commercial aspects of pricing and marketing of MIDI systems are just beginning to permit research and development into the large scale use of single-user multi-tasking 16-bit or 32-bit microcomputer systems which would be ideally suited to the integrated music system environment. The real point here is that so far computers have entered into the musician's consciousness only in certain restricted sets of circumstances. Buying a microcomputer system, a MIDI interface and some software is an alternative route towards the acquisition of a MIDI sequencer. Buying a musical micro such as the Yamaha CX5 provides a sound source and MIDI control in one package. Hopefully, wherever a computer is employed in such systems there will be future expandability via software updates which should be cheaper than replacing the old kit with new. The versatility of computer systems means that they will provide a great deal of functionality including sequencing, patch storage and editing, soft control over the connection of musical kit, and the filtering of MIDI control information (which I will talk more about later). So, unlikely as it may seem, many musicians will see a computer system, however heavily disguised, as the central component in their musical life!
At this point it would be very tempting to dive off into the deep end and discuss the MIDI interface hardware, the communications protocol that goes with it and the detailed ramifications thereof. For this you will have to wait for later articles. In order that we all start on common ground, it is worth clarifying exactly where MIDI fits into the modern synthesizer. This section is certainly not going to be too intellectually challenging but should at least clear up the odd bit of confusion that some musicians will have been shy of admitting to. The fact is that the manuals that come with your latest acquisition often assume a patchwork of knowledge about MIDI on the part of the purchaser and sometimes the blanks have been left unfilled!
The flow of control information in a typical MIDI equipped synthesizer is shown in Figure 2. Note that the instrument has been logically sawn in two so that the sound production facilities are separated from the keyboard and controllers. In fact, many manufacturers effectively (but very neatly!) 'saw' their keyboard synthesizers in half in order to provide the mother keyboard/expander combination and since the relationship between controllers and sound sources is the same in each case, I have labelled the sound source "Expander" in Figure 2.
Although we might talk of parts of the synthesizer as being logically distinct or discuss various logical routes down which control information flows, it is likely that one or more microprocessors actually provide the facilities and handle the control information. Typically, one microprocessor might deal with scanning the keyboard and controllers and transmitting/receiving MIDI, whilst a second actually performs the mathematics used to model waveforms and simulate envelope generators. Eventually, the results of the calculations are converted into sound by a digital-to-analogue converter. On other synthesizers the oscillators, filters and amplifiers may physically exist but their control signals still come via microprocessor control.
In Figure 2 we are looking at the logical connections and internal configuration when the synthesizer is being played by its own keyboard and controllers. The result of any activity on the keyboard or controllers will be a stream of control information which will be fed to the "Expander" and also to MIDI Out so that other synthesizers maybe controlled by and, in fact, mimic the activity. Any control information coming from outside the synthesizer can be fed into MIDI In (and then ever onward via MIDI Thru) when it will be merged with the internal control flow. Thus, it is possible with an 8-note polyphonic synthesizer to have a sequencer play 4 notes, say, via MIDI In, and for you to play another 4 notes via the keyboard.
The 'switch' drawn in the direct internal route from keyboard to "Expander" is implemented in software rather than hardware and is normally closed (or non-existent!). It appears on the diagram to draw your attention to a MIDI feature which is unfortunately not implemented on quite a few synthesizers. It should be possible to tell the synthesizer, via MIDI, to break the direct local connection. You would then have a sort of mother keyboard and expander combination quite literally - but in one box!
If the synthesizer's MIDI Out, which is still outputting keyboard and controller activity, is connected back into the MIDI In via a computer system then all sorts of filtering becomes possible. This subject will be covered in more detail in later articles when the actual filtering of the MIDI data stream will be explained. As an example assume that a second synthesizer was to sound the chords played on this keyboard and this synthesizer was to play only the top note of the chords as a monophonic lead line - the computer could investigate the MIDI control codes 'on the fly' and only send the top note back to the "Expander" part of this synthesizer.
Figure 2 is concerned with the synthesizer in its general mode responding to control information from its own and other MIDI controllers. Note that the internal communication need not necessarily use MIDI protocols - in fact, this would be quite inefficient. Provided the MIDI protocol is understood at the MIDI In and transmitted correctly at MIDI Out and Thru, all will be well. In Figure 3 an additional control path (from the "Expander" to MIDI Out) is shown which allows the "Expander" part of the system to communicate with the external world as well as being controlled.
The normal performance related MIDI control information is not specific to one synthesizer type or a particular manufacturer - if a key is played on a controlling keyboard then all slave synthesizers should understand that control and respond in a suitable manner. This is not the case where, for example, patch information for a particular synthesizer is concerned since a different synthesizer with different facilities, configuration and even method of synthesis, would have no chance of understanding any patch information from a different family of synthesizers.
Where a family of synthesizers has the same internal structure, as in the case of the Yamaha DX7 and TX7 expander (or where one's structure is a subset of the other's eg. the DX9 and DX7), it is possible to pass patches, possibly modified by a computer, back and forth in MIDI's System Exclusive mode. Here, it is the data from the "Expander" sections of the synthesizers, rather than the controller-based performance information, that is being transmitted and received. It is possible, therefore, to edit a voice on one synthesizer and then load it into an expander unit which has no editing facilities. Another useful possibility, once a computer system is connected to the "Expander" section, is to store complete libraries of voice patches, in some sort of database, on the computer's disk storage far more cheaply than on the RAM packs supplied by the manufacturer.
One of the effects of digital control internally of the "Expander" is that modern synthesizers tend to have far less knobs and switches on the front panel (they cost money!). A small display tells you what parameter you are editing at the time and a data entry slider lets you alter the value. This restricted view of the machine's internal state can be quite mind-boggling at times and comparing the 8-stage curves of two envelope generators can be a difficult task! The computer system comes to the rescue yet again by displaying and letting you alter related sets of parameters which it can read from the "Expander", manipulate and then transmit back.
This mode of operation also permits the creation of far more complex and versatile sequencer systems and performance controllers than are in general use today, since it is possible to control every internal parameter remotely. If you want to do something a little unusual, such as change the effect of a controller from linear to exponential at a particular point, it may be possible to connect a pedal to control the change-over in that way during performance. On the Oberheim Matrix synthesizers, for example, you can just about connect anything to anything else! However, if the designers of your synthesizer hadn't thought you might want to use such a form of control it might not actually be possible to set it up from the instrument's front panel. With the magic computer system you can monitor MIDI Out for the pedal being pressed and then change the parameter with a System Exclusive message sent to MIDI In!
As the use of MIDI quickly evolves, the computer power that is obviously a central issue pops up in many different guises. The Mother Keyboard concept reflects the fact that controllers and sound sources are easily separated and can communicate quite complex sequences of control remotely. Since there is no need to purchase a redundant keyboard with each synthesizer why not purchase one high quality and very versatile central keyboard controller unit?
Such keyboards tend to have weighted, piano-type keys which are velocity and after-touch sensitive. The onboard microprocessor can easily handle split points and overlaid keyboard sections as well as complex logical routing via different MIDI channels. Figure 4 shows one physical keyboard logically split three ways. Note that just one key is allocated to setting off the sampler and keyboard velocity could well have been assigned effectively to the size in megatons of the bomb! The other logical keyboard sections may well have their control data musically transposed and their MIDI channel changed etc before appearing at MIDI Out and it is quite possible for both expanders to respond if keys in the overlapped range are played.
An important point when buying a mother keyboard is that the performance controls, such as pitch and modulation wheels, switches and sliders, must be able to drive whatever expander modules you are likely to connect up in the future. The pitch wheel, for example, has a specific MIDI identity (controller number) which both controllers and expanders should all know about, but the identity of other controllers such as the DX7 breath controller, is not laid down in the MIDI specification and is therefore more arbitrary!
So, it should be possible to tell the mother keyboard what MIDI controller identity to programme in relation to its own physical controllers in order to map the keyboard velocity from a Yamaha mother keyboard onto the filter control of a Roland expander, for example. There should also be the provision of 'soft' controllers which can send out arbitrary System Exclusive messages.
Hopefully, having read this article, you may have a clearer understanding of what MIDI can do for you. The point I have emphasised throughout the article is that it is computer power which supports MIDI and has allowed all the very useful interconnections of musical devices. So far, we have only scratched the surface of the MIDI system's capabilities by allowing dedicated control systems such as sequencers to do useful work for us but purely as one system component.
The next stage in system development is to introduce more powerful multi-tasking computers as the central component of our MIDI world. The system integration and enormous functionality provided will be made extremely accessible and usable by the 'musician specific' user interface. Various streams of work will be able to go on in parallel such as printing out a musical score whilst editing another at the same time. Eventually, the Fairlight for the common man may be some hardware and software add-ons for just such a system!
As is the case with describing any complex system, it would be much easier to understand the overview if you already understood the internals and vice versa. This sort of circular learning path is therefore fraught with difficulties but at the end of the day everything should fall into place. So, if some of the concepts discussed in this article look useful but you can't quite see how MIDI helps or if you can't see why some of the ideas described might be useful, don't give up hope yet! As the protocols used are explained and more concrete examples are described in later articles, all should hopefully become clear.
Readers please note: if you are having problems with any aspect of MIDI, drop us a line and let SOS help.
Feature by Jay Chapman
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!