Much has been said about MIDI patchbays and their near indispensability in a modern MIDI setup. Vic Lennard questions their importance and looks at the alternatives.
MIDI PATCHBAYS ARE CERTAINLY A CONVENIENT WAY OF MANAGING MIDI EQUIPMENT, BUT ARE THEY ESSENTIAL - OR ARE THERE CHEAPER ALTERNATIVES?
THERE COMES A point in the setting up of a MIDI system where a decision has to be taken as to how the MIDI signals are going to be routed. The most inviting of devices is the all-singing, all-dancing MIDI patchbay which will do everything bar making the mandatory cups of black coffee (no sugar).
A MIDI patchbay is a matrix of MIDI inputs and outputs which allow you to patch between them by pressing a few buttons. The more expensive ones also incorporate a degree of processing - possibly including rerouting according to MIDI channel or note velocity, filtering various elements out from the input and remapping MIDI controllers.
But is this really the best device for you? To answer that you need to decide exactly what the MIDI side of the system has to be capable of achieving, so let's have a look at a few typical setups.
THE MOST COMMON type of system is one in which you have a sequencer sending MIDI information to various MIDI modules. There may be a synth which is acting as a master keyboard and using local off to divorce the keyboard part from the sound module. This is connected via two MIDI cables to the MIDI In and Out of the sequencer. Other modules are then connected MIDI Thru to MIDI In in a "daisy-chain". But why bother using another piece of hardware when almost all your devices have MIDI Thrus? There are three reasons: firstly, a MIDI Thru will not operate unless the device containing it is turned on. (Obvious perhaps, but this means that all the devices in your system are going to have to be on all of the time.) Secondly, the opto-isolators in the MIDI interface age and so slow down. This doesn't lead to delays (as popularly believed) but will affect the shape of the waveform from the MIDI signal - instead of being a square wave, the leading edge starts to slope. Thirdly, if you want to use your computer as a librarian - that is, for transferring sounds to and from the computer using MIDI System Exclusive messages - you'll continually need to repatch your system.
So we need a MIDI patchbay, right? Wrong; all you need to deal with the first and second situations is a MIDI Thru box which takes a single MIDI In and turns it into many Thrus. Typically this costs between £15 and £35, depending on the number of Thrus that you require. We'll look at the bulk librarian issue later.
UP A STEP in sophistication from the previous example is the situation where you have more than one MIDI master controller. These might include keyboards with weighted and unweighted action, a drum machine, a guitar synth and a wind controller. On some occasions you might want to input notes to the sequencer from a keyboard, on others you'll be tapping away on the drum machine pads or blowing soulfully on a MIDI sax. In order to be able to do this, you either have to repatch between controllers or use a MIDI patchbay, right? Again, wrong. All you need in this case is a MIDI switch box. This usually has a number of MIDI Ins and one MIDI Out; by rotating a switch you can select which of the Ins feeds the Out and, hence, which controller is the current master. Of course, you can't switch while MIDI data is passing through as you run the very real risk of data corruption, but a switch box will remove the necessity of plugging and unplugging MIDI cables.
Cost again is quite low - typically £25-£40. There are also MIDI switchers with two MIDI Outs so that you can have two controllers patched to different places at the same time.
AS MENTIONED EARLIER, a MIDI patchbay will often have an in-built processor for altering MIDI data as it passes through the unit. So we need a MIDI patchbay for processing data, right? Still wrong, I'm afraid. If you're working with a sequencer, and most of you reading this will be, the facilities for data processing on the sequencer are likely to be quite comprehensive. The most obvious example of data processing is input filtering, where unwanted MIDI information can be removed from the incoming data stream. This might be useful, if you're playing drums from a keyboard with aftertouch sensitivity - aftertouch won't add anything to the drum part, and removing it will save sequencer memory. As another example you might filter out All-Notes-Off messages - this saves the awkward situation where you record, say, a piano in two takes with the left hand first and the right hand second. The second take will have notes cutting out whenever you removed your fingers from the keys in the first take.
Most sequencers will also let you remap MIDI controllers. If you're using a Yamaha WX7 wind controller, the loudness of each note is controlled by the lip movement on the sensor (sent out as MIDI controller #2, Breath Control). Apart from Yamaha equipment, not many other synths respond to this controller but you can get the same effect by remapping this to MIDI controller #7, MIDI Volume. As the relevant data is received at the MIDI In, it is altered into its new form before being recorded.
USING A VISUAL editor for a modern synth makes patch editing a much more appealing prospect. But what happens if that synth is a module - so it has no keyboard of its own - how do you edit it and play it from an external keyboard at the same time? Time yet for the MIDI patchbay? Not quite - instead, a MIDI Merge box will merge the data from both MIDI Outs and send it to the module. MIDI merging requires a processor because MIDI messages are usually made up of two or three bytes which have to stay together if the message is to be correctly understood. Cheap MIDI patchbays generally don't have onboard processing and so cannot carry out merging, while a dedicated merger can be bought for less than £100.
SO WHEN, IF ever, do you need to consider using a MIDI patchbay? The most obvious application is when you have many MIDI master devices and modules and so would need a maze of MIDI Thru and switch boxes along with a merger to cope with all eventualities. A MIDI patchbay would solve the problem that the cabling would present. But do you need to spend several hundreds of pounds on a sophisticated unit?
To answer this, a little MIDI patchbay theory is in order: most units that are moderately priced are software-based, in that MIDI data arrives at an input controlled by a Universal Asynchronous Receiver Transmitter (UART) which informs the processor of this occurrence. The byte(s) are then stored in RAM, altered in whatever way is programmed and then output. The problem with this is that if too much data appears at the inputs at the same time, the system becomes overloaded. This leads to delays (on a good unit) or corrupted and lost data (on a bad one). A well-written software program is essential for the efficient working of a MIDI patchbay and even with this, there has to be a minimum delay in the moving of data to and from RAM. The moral is if you don't need processing power in a MIDI patchbay, keep to a simple design which routes Ins to Outs for minimal delay of MIDI data.
A non-processing patchbay with four or eight Ins going to eight Outs will cost around £100-£150 secondhand. Units like the Yamaha MJC8 and Akai ME30P are quite easy to find secondhand. There is also the Philip Rees 5X5 which is a five-input, five-output unit operated via front panel switches and which can be expanded.
The simple addition of a merger makes the unit a little more expensive but you should be able to find an Akai ME30PII or a Roland A880 in MT's Readers' Classifieds for under £200.
A company called DACS had a rather clever idea on MIDI patching: as there are only two active connections in a MIDI lead, they designed a MIDI patchbay similar to its audio equivalent. This meant that you could set up patches with ¼" jack-to-jack leads on the front panel. The unit also included the ability to hard wire a preset configuration of the eight inputs and outputs. It was a little restrictive in complicated situations, but represented an inexpensive patching system.
Isopatch and MTR have passive patchbays which require you to use short MIDI leads for patching. These make life easier in that all the patching is done in one place (again like an audio patchbay).
ONE OF THE pleasures of having a sequencer with a disk drive is being able to save sounds to disk - like 180 DX7 banks onto a DS/DD disk. But how do you get the bank into the sequencer? Either you initiate the dump from the synth or send a request command from the sequencer - so you're going to need a two-way MIDI connection allowing you to send sounds between two. This isn't too much hassle when you only have one or two synths (use a switch box), but when the system grows and you have perhaps six synths, switching between them each time you want to transfer sounds is going to get tedious.
One solution is offered by the Anatek Studio Merge - an eight-way merger with corresponding MIDI Ins and Thrus - by using this and a separate Thru box you can set up eight modules, each with their own handshaking MIDI configuration. Not cheap, but if that's the way your system functions it is difficult to conceive of a simpler setup.
BACK TO SQUARE one. Some people simply put up with repatching MIDI leads while others ignore the cheaper options and buy the latest fad in MIDI patchbay technology. Part of the decision will also depend on your sequencer: some hardware sequencers have multiple MIDI Outs, while the manufacturers of software sequencers often have their own proprietary interfaces giving extra MIDI Ins and multiple independent MIDI Outs.
I can offer one last piece of advice, in your quest for the ideal patching system. Before you go out and buy MIDI routing devices, work out how many inputs and outputs you need for your current system - and double them! Failure to take expansion into consideration will invariably result in costing you more money when you have to update inadequate devices a few months later.