Stars & Chains
Martin Russ outlines the practical benefits that using a Star Network can bring to anyone's existing MIDI set-up and further explores some applications of MIDI Switching Units.
Want to get more out of your present MIDI equipment but don't know what to connect where? Contributor Martin Russ lays down some simple practical guidelines which, if followed correctly, should increase the creative scope of your MIDI set-up.
Devices to route MIDI signals have recently appeared from several manufacturers, offering the user quick and easy changes to MIDI-based equipment set-ups, without any need to invest in computer-based control systems. These are known as MIDI Switching Units (MSUs) and can vary from the very simple, eg.the Skyslip MIDIswitch - suitable for the smallest home studio, through the application-specific Roland MPU-104, MPU-105, and the versatile Yamaha YME-8, right up to sophisticated professional rack mounted versions like the Sycologic M16 MIDI Matrix. This article describes some possible applications for these units, and highlights a few unusual and unexpected side-effects you need to be wary of.
Most MIDI-based pieces of equipment come fitted with the bare minimum of MIDI interfacing: on a synthesizer it is usually just a MIDI In and Out DIN socket, with perhaps a Thru for good measure. Some MIDI-controlled effects units have only an In, with nowhere for the data to escape from! The number or type of inputs and outputs is not really important if you only have a couple of pieces of equipment to connect together (although you can use MSUs to advantage in these circumstances as well), and it is really only when you have more than three MIDI items that it becomes important to consider how to interconnect them (and how to change the configuration later). Once you begin to join several MIDI-based units together to form a Network, several unexpected things start to become very important, as we shall discover.
There are two main ways of connecting lots of anythings together, be they computers, chained up prisoners or copies of this magazine: Stars and Chains.
Stars require a central controller, often called a 'Mother' or 'Master Keyboard' in MIDI set-ups, and all the other pieces of equipment are connected directly to it. Thus a separate cable connects each peripheral unit to the central Master Controller. See Figure 1.
A Star Network thus has a Master Controller at the centre, with peripherals connected as Slaves, each with their own individual cable. Each Slave can be referred to as a Branch of the Star. MIDI Outs from the Master are connected to MIDI Ins on the peripherals. Because of this, Networks configured as a Star have the disadvantage that the Master Controller needs lots of MIDI Outs; but they also have two distinct advantages:
1. They minimise the time delay when transmitting from the central controller to any peripheral unit.
2. There is only a single connection between the centre and the peripheral, making switching and control easy.
An example of a Star Network is your local telephone exchange - all the local phones have separate lines to the exchange, where all the switching of local calls occurs. Thus when you phone your neighbour, your call goes to the exchange and then all the way back to the neighbour's house - just to get next door! The central controller in a Star Network can control all the other units, if so desired, but as you can see, complications arise if you want to use another keyboard as the Master. To change controllers you need to reconnect all the MIDI Outs from the old controller to the new one - not very convenient...
Chains also have a Master Controller, and this could also be a Mother Keyboard, but in this case the other pieces of equipment are connected serially through their MIDI Thru socket (hence the name!) ie. linking them in a chain (Figure 2).
Chain Networks have the advantage that they can be easy to set up if your equipment has the necessary Thru connections, as well as enabling you to use any keyboard in the Chain as a controller for whatever units are connected after it, by switching between the Thru and the Out with a MIDI Switching Unit. (Remember that the first controlling device always uses its MIDI Out to connect to the rest of the Thru'd chain of instruments.)
In a Chain, as in a Star, it is possible for the Master unit to control all the other pieces of equipment in the Chain by setting all the units to operate on the same MIDI channel (or you could just set them all to OMNI mode, but this obviates any separate control of individual units). Unfortunately, there is often a time delay between the receipt of information at the MIDI In port, and its subsequent re-transmission via the MIDI Thru.
These time delays also increase as the length of the Chain increases, with the result that supposedly simultaneous notes are turned into an (unwanted) arpeggio, or more disastrously, the sharp, digital attacks of notes are smeared into an unimpressive 'mush'!
At some point in your connection experiments, you may feel tempted to try taking the last unit's Thru and connecting it to the Master's In. This looks like it should form a loop, where MIDI signals can travel around and around without hindrance - happily it is quite difficult in practice to create a loop. Look again and you will see that because the Master controller unit uses its Out to control the later units, it is not possible to produce any MIDI information at its Out, except by generating the information from the Master controller. This is so because the MIDI information coming into an In is repeated only at the Thru, not at the Out. (The more observant may care to note that this has not always been strictly true - one early MIDI version of a well-known drum machine repeated everything presented at its In, on its Out - instant loops for the unwary!)
On a similar theme, you may also like to try producing a Chain by connecting the Outs of your MIDI equipment to the Ins, instead of the Thrus to the Ins - this has the unfortunate result of producing a Chain where each unit can only control the next one down the line-possibly the least useful of all MIDI networks. Try it if you have lots of hands....
Now although the MIDI specification allows for both Star and Chain Networking, the time delays and lack of Thrus needed for a Chain tend to make the Star the usual choice.
Unfortunately, Star Networks need more than the single MIDI Out on the Master Controller (Figure 3).
In order to solve this, the MIDI Thru Box was invented. It acts like a signal splitter box ie. it turns a single Out into lots of Thrus, so you can wire up a Star Network which has the minimum of time delays, and you have only a single Master keyboard to deal with. A simple Star Network would thus look like Figure 4.
The simplest mode of use of such a Network is to set all the units to OMNI, whereby everything you play on the Master Keyboard is also played by the other units. That's fine for fat, chorused sounds, but not very good if you want multi-part orchestrations. A more practical and flexible approach is to set all the MIDI channels to be the same and thus all units will play the same notes as the Master keyboard - since some MIDI instruments have restricted channel selection, the channel chosen is normally 1. Separate control of the individual component units can then still be accomplished by using the remaining 15 MIDI channels and assigning each unit to a different channel. Eagle-eyed readers will already have spotted that if all the instruments are set to channel 1 as above, then they can't also receive data on another channel destined exclusively for them. Very often you would also like to hear just one unit instead of all of them. To solve these sorts of problems you need either a MIDI Switching Unit (MSU) or a MIDI 'Or' Box.
MIDI Or Boxes change MIDI channel information so that incoming channel 1 data is either passed through the device unchanged, or changed to the desired channel. This needs quite a bit of electrickery to achieve and such devices are suitably expensive to purchase (eg. Dynacord MCC-1). In their defence, they can be used without recourse to complicated wiring since they can just be connected in the cable leading to the unit in question.
MSUs (MIDI Switching Units), on the other hand, vary from the cheap and simple, to the expensive and adaptable. Their use usually requires a little bit of careful forethought, and a bit of extra cabling. They have the advantage of being very flexible in their applications: you can thus choose a configuration to suit your own requirements, and expand it as required. For solving the problem mentioned above, specifically that of choosing which units play at any particular time - instead of selecting by using the channels, with MSUs you choose which peripherals you want instead, leaving you the option of using different MIDI channels for further control.
Let's look at a low budget example consisting of:
1 Master Keyboard
1 Sequencer or Computer
1 Expander/Slave Synthesizer
With all three pieces of equipment set to MIDI channel 1, we face a dilemma - how is it possible to control the Expander from the Master Keyboard AND from the Sequencer? Well, by using a simple 2-in/1-out MSU, for example the Skyslip MIDIswitch, it is possible to choose either the Master Keyboard or the Sequencer to act as the controller for the Expander, as well as isolating the Expander from both controllers. (Figure 5.)
If we were using a Slave Synthesizer (functioning as an Expander) in place of the Expander unit, then this isolation case has a further use. With the Master Keyboard connected via an MSU to the Slave Synthesizer, notes played on the Master will also be played by the Slave. But if a note or notes are played on the Master and the Slave is isolated while the notes are held, then the Slave Synthesizer will continue to play the notes, even when they are released on the Master Keyboard. This is the cause of the infamous 'sustaining note' phenomenon; if a MIDI connection fails in mid-performance, you may find it very difficult to stop the music...
With a foot-activated MSU, you could use this effect usefully to sustain chords without using your hands. Since most Expander-only units utilise MIDI's 'Active Sensing' function to detect the loss of a controlling Master Keyboard (which automatically shuts off the offending note or chord being held), this sustaining effect will only work with synthesizers used as expanders (that CZ-101 looks a good investment, after all!). This is a little-publicised advantage of buying synthesizers with keyboards - like the Alpha Juno-1 and DX100 (however small and restricting) - rather than dedicated expanders which do nothing but sulk when isolated. Having said that, I also sulk when isolated!
For larger and more complicated MIDI set-ups, a slightly different approach to Networks can be very useful. Instead of connecting all the peripheral units to the Master Keyboard in a Star fashion, you only connect the units which are time-critical - usually synthesizers. This leaves the Thru and Out sockets on the starred units available for the connection of further Slave Synthesizers or even the MIDI-controlled effects units which are now threatening to become ubiquitous - the obvious example being the Yamaha SPX-90. A Network connected in such a way is known as a Tree. (For the curious - this explains why I mentioned the Branch notation before, when talking about Stars.)
Tree Networks have a Master Controller at the bottom, forming the Tree-trunk (with the mains connection presumably providing the roots and earth connection!). The peripheral MIDI units (Leaves?) are connected via Branches to it and less critical units can be connected as Stars or Chains, via Branches (Twigs?) from the first set of units. See Figure 6.
As an example: for a synth and MIDI-controlled effects unit used together, you can chain the effects unit and the synth it treats (connected as part of a Star Network to a Master Keyboard) via a special connection of two simple MSUs, thus providing overall control via the Master Keyboard, or local control of the isolated Branch. The 'special connection' is called a Localiser, and is another MIDI box which enables you to change over from a Slave to an isolated local Master, and is especially useful when used in conjunction with an effects unit, as here. You can then use a single Localiser to locally override the central controller (Figure 7).
In many cases, a small and simple MSU like a Localiser can have the advantage that it is in close proximity to the equipment in question, as well as being easy to alter and add to - unlike a large, central Switching Matrix, which may need a large number of long cables to connect it up, and may make it less obvious what is going on. A useful technique to adopt here is the use of a thorough cable marking system. A few bands of coloured heatshrink or coloured tape at each end of a cable, or several coloured dots near a socket, can make living in a cable-strewn jungle a lot less venomous.
Having said that, a central MIDI Switching Matrix (if you can afford one) has several advantages. It removes the need for Thru boxes and Localisers, since the functions of each of these can be achieved within the Matrix.
Software control of the Matrix can also help you to rapidly reconfigure the network, where the use of simple MSUs would mean elaborate, time-consuming, unwieldy and downright confusing connections.
The use of a Switching Matrix, however, still depends on knowing and understanding the basic ideas - unless you knew what a Localiser was, you might never think of using one. And once you familiarise yourself with the basics, you should be able to cope with just about any MIDI Switching Unit (MSU), regardless of how large and awesome it may appear, by considering it as lots of smaller and simpler MSUs. You can learn a great deal by using a couple of basic MSUs first, before you plunge into a large Matrix (Figure 8).
The effect of using large MIDI networks on the user can be pleasurable or distressing. Hopefully this article will help guide you through to ultimate success with MIDI.