Sixteen MIDI channels not enough for you? It could be time to find another way Out...

When MIDI was first developed some ten years ago, instruments were generally monotimbral: that is, they could play only one sound at a time from MIDI notes received on a single MIDI channel. So the 16 available MIDI channels could control up to 16 different synths or sound modules. But of course, the technology has developed to such an extent that a single multitimbral sound module may now require all 16 MIDI channels, particularly where such units boast a polyphony of up to 64 notes.

Take a typical live set-up of a synth and a couple of sound modules cabled together as in Figure 1.

Playing keys on the synth sends out note information to both of the modules. But let's say that each of them is 16-part multi-timbral - General MIDI modules, perhaps. The only way this system can operate is with both modules in tandem, responding to the same note information on the same MIDI channels. Also, a Program Change message sent from the synth will switch both modules to the same patch number, so you'll have to carefully organise your sounds if this is not desirable.

Now, what happens if your synth has a pair of independent MIDI Outs? The set-up would be as in Figure 2.

Here, each sound module can be individually addressed from the keyboard and the problem with patch changing no longer exists. Such a synth is likely to support multiple splits so that different zones of the keyboard transmit MIDI information on discrete MIDI channels - a very powerful live system.

Working live with a sequencer/workstation or back in the studio with a sequencer, there's a much more important reason for having independent MIDI Outs. Taking the above example of two 16-part multitimbral sound modules: having one MIDI Out means only being able to access 16 instruments. While this might seem a lot, it doesn't allow you to individually use each sound on both modules.

The reason for having a pair of expanders might be to make use of their different synthesis methods. For instance, one might be of the sample playback type offering 'real' instruments like strings, brass and the like. The other unit may be a standard synth for pads, allowing you to double up sounds to create useful textures. Additionally, being able to automatically access all instruments on a pair of sound modules means you are far less dependent on patch changing.

Figure 3 shows a typical set-up for this; here, one MIDI Out plays the sounds on the synth while the other is used purely for the two sound modules, using the MIDI Thru of the first to connect to the MIDI In of the second.

A second MIDI Out can also be used for the sending of accurate sync information. Obviously, the timing of this kind of data is very important, bearing in mind that it is responsible for keeping two devices locked together - a sequencer and a drum machine, for example. This is shown in Figure 4.

MIDI Out1 transmits the note and performance information to the synth while MIDI Out2 is responsible for sending the MIDI Start, Clock, Continue and Stop commands to the drum machine whose patterns will then play in time with the sequencer. While the speed of MIDI is not usually an issue, a sequencer will often prioritise the transmitted MIDI messages, and MIDI sync commands generally come way down the list. By using the second MIDI Out in this way, you ensure that the playback timing of the drum machine is not compromised.

While generic multiple MIDI Out units are available for the PC and Mac, sequencer software manufacturers for the ST tend to provide dedicated units. These often provide you with more than simply an extra MIDI Out - key (dongle) expanders, SMPTE generators etc. are also included. For most ST sequencers, there are cheap third party options such as 16 Plus from Hands On MIDI Software and ModemMIDI from the UKMA.

General MIDI has been mentioned in the main text - but what exactly is it?

General MIDI

Suppose you use a particular sound module with a sequencer to record a song, and on playback, switch to a different unit. For the song to bear any resemblance to the original, the following have to be true:

1 Any Program Change numbers used must call up the same type of sound on the new module.

2 The MIDI Note numbers assigned to the drum and percussion instruments have to be the same - and on the same MIDI channel.

3 The polyphony of the new module must be at least as high as the original one.

4 The new module must respond to any MIDI Control Changes that have been used, such as MIDI Volume and Modulation.

General MIDI was designed with all the above criteria in mind. All GM sound modules have the same basic 128 sound types assigned to the same MIDI Program Change numbers according to the GM Sound Set. This means that if you have selected Program Change #1 (Acoustic Piano) when recording, the same patch (or a sound close to it) will be called up on playback. Any key-based percussion is always assigned to MIDI channel 10 and follows the GM Percussion Map which is, perhaps, a little limiting as only 47 sounds are used - 20 of which are Latin percussion instruments - with only two bass drums and a single acoustic snare drum.

In terms of polyphony, any GM module must provide at least 24 voices - although, as the specification currently stands, more than one of these may be used per sound, thus making it impossible to guarantee a specific number of notes. Finally, certain MIDI Control Changes must be recognised, including Modulation, Volume, Pan, Expression and Sustain Pedal.

How do you know if a sound module is GM compatible? Look for the logo (Figure 5). But be careful: some modules are using the logo without properly conforming to the specification. Check the polyphony - if it can't provide 24 voices, it isn't GM. The result of this will be the audible cutting off of notes.