Regular readers and newcomers with a well developed sense of humour, will know that each month, on pages 8 and 9, we have 'a funny' — a few hundred words dedicated to the idea that musicians enjoy a laugh (the OTT theory) and don't spend all their time as miserable as sin, poring over voice coil specifications of the latest loudspeaker (the rest of the magazines).

June's concerned MIDI, which, as someone later pointed out, is no giggling matter. Very true. But neither should it be delivered into the hands of the techno-prat who implies that you can't use it, unless you can recite the software structure of your synth backwards in your sleep.

So, in traditional One Two fashion, we take a look at MIDI, at age two — offer some definitions, explanations and experiences, and do it for the average, everyday musician; That's you.

It stands, as we all know, for Musical Instrument Digital Interface, was launched in '83 (first appearing on Sequential's Prophet 600 and Roland's Jupiter 6), and came about as the result of discussions between most of the major synth manufacturers searching for a common form of interconnection.

In its most basic form it allows one synthesiser to play another; at more refined levels, it can for example, change voice programs at the instruction of, say, a MIDI equipped sequencer, swap program data, or handle instructions from a computer.

Information is transmitted serially (one bit after another in a straight line, rather than eight bits presenting themselves side by side), and is asynchronous... all connected items will act on instructions individually, as fast as they can. The opposite of this arrangement would be a system that wouldn't start any job until every component — including the slowest — was ready to go. You have the advantage of each unit working at its own best possible speed, but some timing disadvantages can arise, as we'll see later.

The speed at which this information travels is the Baud Rate and is fairly high compared to personal micros. Only two out of five pins are used on the Din connector (selected for availability and low price), and there is no direct electrical connection between the incoming info and the receiver. It's translated into flashes on an LED then read by a phototransistor. The lack of direct contact avoids hum loops.

Most of the above regimes were decided on to make MIDI relatively affordable and easy to deal with so that all synth makers could incorporate it. Even those at first reluctant have now come round. Interesting to note that companies who once spurned MIDI used to argue that it was impossible to fit MIDI to an existing machine.

"Too complicated," they said, "you'd have to rebuild everything from the ground up." Today MIDI retrofits are not uncommon.

Even now the system is not without its quirks, but it still works better than anything keyboard players have had before. However, compromises create problems, and the major drawback MIDI faces is one of speed.

It's estimated that it takes a synth a quarter of a second to react to MIDI info telling it to 'turn the note on'. In comparison, some envelope generators can produce attack times comfortably below 10ms. So 250ms is not an inconsiderable delay in feel.

Being a serial system, everything takes its place in the queue. As MIDI is capable of 16 note polyphony, on 16 different keyboards, every note of a 16 part chord will come down the line one after the other. If they pop out at a rate of one note 'on' per 1ms, the last will be heard 16ms after the first. Feed that to a delay line, and you'd be in double track territory heading for slapback echo. Fortunately few of us have 16 fingers and practically (four or five part chords) you'd need a sharp ear to spot the problem.

MIDI also adds to the workload that the internal microprocessor of your keyboard needs to carry out. The busier it is, the greater the chance of hesitation. Likewise, the more MIDI keyboards you have connected together to handshake the data along, then the longer the synth at the end of the pier will have to wait for its splash of the water.

However, every cloud, etc, and I've read of keyboard players who say the delay can produce a chorusey ensemble feel to racked up MIDI synths.

One last cautionary chapter about information dispersal. Chaining.

On the back of most newer MIDI keyboards you'll see three sockets — In, Out and Thru (originally there were just In and Out). Typically, transmitted data from the Out of one synth is connected to the In of a second, and 'echoed' back from Thru, to the In of a third, from Thru again to the In of a fourth, and so on. You don't use the Out socket of each synth to drive the next, because it's there to transmit only the MIDI material that particular keyboard is generating.

Of course, there's a snag. The echoed data comes after the LED and phototransistor (together known as an opto-isolator) have done their business. By the time the fourth synth gets its instructions, the original messages have been through this process three times. Just like constant tape copying it becomes distorted. The 'edges' are knocked off and the data can no longer be reliably interpreted.

Reports of this phenomenon vary, but you ought to be safe with three outboard units. After that, it will depend on the quality of opto-isolator used by the builder of your set of ivories.

One way round it is to use a MIDI thru box which can send a number of parallel signals to connected synths, all from one MIDI in.

There are several to be found on the shelves, usually offering four or six outputs.

Alerted to the limitations, we turn to the lexicon of MIDI which is especially fond of the word Mode, fast becoming to the English Language what 'situation' was in the early 80's.

MIDI info can be assigned to 16 channels. In the earliest format 'Omni' meant the synth responded to all data hitting it regardless of channel (it's the standard mode a MIDI device will default to when first turned on). Omni is now split into Poly and Mono, operating as you'd expect.

In the alternative 'Omni-off' mode you can specify a channel polyphonically or individual voices monophonically, the latter technique developed for synths such as Sequential's Multi-trak where each of the six voices can have its own sound and its own MIDI channel.

As interconnections grow more complex, you may not want all of your instructions from the master keyboard going down the line. Synths with a sophisticated MIDI mentality can be asked to ignore, say, pitch bend or mod wheel datum when it reaches them. Such transmissions, along with patch change and velocity details, etc, are lumped together as 'Channel Voice' subjects. 'Channel Mode' contains the Omni on/off instructions, 'Real Time' is the channel wide MIDI 'beat' that runs drums machines and sequencers ('Common' messages cover the rest of the info for these devices) and 'Exclusive' is where the individual manufacturers go to the Bahamas.

The Exclusive realm is where manufacturers corral operations specific to their own keyboards which generally cannot be accurately interpreted by other maker's synths. In the majority of cases it may even be between synths of the same model number. For example, a Yamaha DX7 and TX7 can swap patches happily, but a Roland Juno 106 connected to either would find the data unintelligible. The Exclusive section of MIDI may also be reserved for the fine details of conversing with a personal micro.

Older keyboards may not be as MIDI-ed as you may think. The earliest synths only have Omni Mode and will attempt to play all 16 channel's worth of material coming to them. If you had such a synth in a MIDI set-up and wanted it to play only one stream of material, you could buy a MIDI Filter to take out a specific channel.

Similarly beware of the MONO mode on older keyboards. It may just be a version of the unison switch on a synth and not offer the multi-timbral facilities of newer devices.

If elaborate MIDI link-ups are going adrift, you can search for solutions in several places. Take out the oldest keyboard — it's the one most likely to latch-up under the ever increasing amount of data flowing down the MIDI line. Likewise, try to orient your machines so that the most sophisticated keyboard (with hopefully the fastest house-keeping micro inside) is tackling the heaviest workload.

Keep the degradation through successive transmissions in mind, and do your best to minimise it by using good, low-capacitance, well screened MIDI cables. If everything appears to check out and you don't believe the material is being distorted by the synths or sequencers themselves, then consider the gremlin of outside interference, especially from the mains. Unstable supplies are bad enough, but if you're working at home under the ever popular fluorescent lights and dimmers switches, these can both put out what is affectionately known as 'hash' which troubles a delicate mind, whether electrical or organic. I'll go along with that.