XRI Systems XR400 MIDI Mate
MIDI Patch bay
Do you need a MIDI patchbay in your studio? Or a MIDI merger? Or both? Vic Lennard discovers that the XR400 solves more problems than he suspected. How's the software-reviewers elbow, Vic?
How far can you develop a MIDI system without needing a MIDI patchbay? Will a MIDI patchbay solve all your MIDI patching problems? If you invest in a MIDI patchbay, should it be the XR400?
MOST OF US with a drum machine and a couple of multitimbral synths in their home studio eventually invest in an audio patchbay, to ease the strain (mental and physical) of continually repatching leads. We could adopt a similar policy to relieve some of our MIDI patching headaches, but there's a problem. The complexity of even a modest MIDI setup seems to be enough to scare us off. Patching a MIDI system is bound to be more complex than its audio counterpart - for starters, MIDI In, Out and Thru sockets are to be found on most devices, which means that you have to work out what equipment is transmitting data to which devices and patch accordingly. The truth is that a couple of synths, a computer with visual editing software and a mother keyboard are quite enough to confuse us all.
But the day finally comes when we've had enough. The only alternative to the asylum is a MIDI patchbay. It must be the solution to our troubles. But hold on, not only do we have to suss out which way round to plug in all the leads but we also have to program the configurations for each situation. Not only that but most patchbay screens are made up solely of numbers and letters (and not in word form). All we really need is some bright spark to suggest a MIDI patchbay as a suitable place to incorporate a merge box, and the asylum starts to sound like an attractive proposition. I was that man.
What's needed is a MIDI patchbay which can accommodate a sufficient number of inputs and outputs, with a merge facility built in that doesn't need to be programmed. XRI, a Birmingham-based company known for their XR300 SMPTE-to-MIDI converter, have designed and built the XR400 MIDI Mate, the first automatic routing MIDI patchbay.
PACKAGED IN A 1U-high 19" rack case, the XR400 has five inputs (A-E) and 11 outputs (1-11). Input C and output three are both on the front panel - good idea - while the remainder constitute the connections on the rear. Inputs D and E and outputs 1 and 2 are specifically intended for a master keyboard and sequencer respectively, while any one of the remaining inputs can be active at one time and are selectable via a front-panel switch. Each input has an associated LED which is on when that input is selected (D and E are permanently on) and which flicker when a data stream is being received. The remainder of the front panel is made up of white labels on which the name of the connected piece of equipment for each input and output can be noted.
THE DESIGNATION OF specific inputs and outputs for the master keyboard and sequencer is intended to prevent MIDI loops from occurring. While this should be sufficient, XRI have also included a software routine which checks the MIDI data at inputs D and E and prevents the same data appearing at their corresponding outputs, 1 and 2.
To understand the idea behind MIDI Mate, it is important to appreciate the internal wiring. The Master keyboard (D) can pass MIDI information to outputs 2-11 while the sequencer (E) passes data through to outputs 1 and 3-11. For both of these, all types of MIDI data will be passed. Inputs A, B and C can pass all data through to outputs 1 and 2 but will have system exclusive information filtered from outputs 3-11. The reason for this is that a MIDI loop could occur when using visual editing software with a MIDI module.
Because a Master keyboard will automatically be connected to sound modules on outputs 3-11, a problem can arise when using a sequencer with software MIDI Thru. The data will reach the modules both from the keyboard input D and via the sequencer input E and will lead to double notes or worse. This "soft" Thru is also used to re-channelise MIDI data. For instance, instead of changing the MIDI channel on the Master keyboard each time a different sound module is to be played, it is easier to simply set the output MIDI channel on the sequencer. To counter this, there is a button labelled ANO (Alternative Note Output) on the front panel, which disconnects the Master keyboard from all outputs except the one for the sequencer, number 2. Consequently the Master controls the sequencer, which then controls everything else. The original design of the XR400 had this as a non-latching switch, which led to confusion as to whether the function was on or off, especially as it defaulted to off when the unit was turned on. That has now been replaced with a latching version. Talking of switches, the mains power supply switch is conspicuous by its absence, and I would certainly prefer to see one.
Should the XR400 lock up in use, there is also a Reset button which sends out all-notes-off and pitch wheel re-centre commands on each MIDI channel. Disconnecting MIDI leads, turning equipment on and off and changing the source of MIDI clock from one device to another are all situations where a press of this button should help to ensure problem-free operation.
ONE OF THE main areas in which MIDI Mate is likely to find itself being used is with a computer sequencer. XRI have apparently used small buffers for MIDI data and rely on the speed of their microprocessor to ensure that these buffers do not overflow and lose important information. A lost Note On means that a note doesn't sound, while a Note Off going astray leads to the note hanging - which will sometimes be sorted out when that particular note is pressed and then released again. Nothing that I did could make the buffers overflow, including running 16 MIDI channels of data at 48Obpm and sending out the equivalent of 5000 notes in the course of two seconds - just the kind of situations that occur in everyday life.
Seriously, no timing glitches or hanging notes occurred, which shows that the theory behind MIDI Mate also works out in practice. In fact, XRI claim a maximum MIDI Thru delay of 300 microseconds which is approximately one third of the time taken to transmit a MIDI note on. The Alternative Note Output switch was permanently on so that MIDI information could be re-channelled. From the point of view of patching, running with a sequencer is no big deal as the MIDI wiring network could be permanently set up if this were to be the only use. However, using the main Master keyboard with another controller, be it drum pads or a guitar synth, requires the use of the merge facility on the XR400, and again it seemed to work fine.
USING A VISUAL editor running on an Atari ST, with a MIDI sound module usually entails one of two possibilities; play the module via the on-screen keyboard (Steinberg, Interval), invisible equivalent (Dr T, Soundbits) or by mouse (Pandora) - none of which can ever give the correct feel and which are often lacking the pitchbend, velocity sensitivity and modulation facilities you might need to audition a sound - or use a merge box.
The necessary connections are for the Master keyboard and sound module outputs to be merged and to control the computer which in turn controls the sound module. Any alterations on screen will be sent to the module and vice versa, while the keyboard sends note data via the computer to the module. MIDI Mate comes into its own here. I tried it with Dr T's editors for the Yamaha TX7 and Roland D110, Interval Music editor for the Kawai K1 and Soundbits editors for the D110 and K1. No problems were encountered with any of these.
"In use, no timing glitches or hanging notes occurred, which shows that the theory behind MIDI Mate also works out in practice."
As far as editors are concerned, the acid test tends to be with the Pandora Technology ones for the Korg M1, D110 and DX7. These are desktop accessories and so co-habit with the sequencer. Using C-Lab's Creator for instance, it is possible to edit a synth while the sequencer is playing it and to record these edits on a separate track (the M1 version was reviewed in MT, December 988). This stretches both the computer and the MIDI patchbay to their limits, due to the mixing of performance and system exclusive data. Using the DX7 editor resulted only in the occasional glitch. The D110 version would have made a more interesting test as two-way MIDI connection ("handshaking") is required, but you never can try out everything, can you? This would essentially be the same situation as described above, with a standard visual editor but with a lot more MIDI note information involved.
THE ONE SITUATION usually guaranteed to cause problems is when two devices are sending out MIDI real time information - such as timing clock and start/stop commands. As soon as the XR400 receives any message pertaining to MIDI timing at an input, that input is designated the "master clock" and will continue to be so until the unit is reset. Any MIDI clocks appearing at another input are ignored. A Roland TR626 was set at input A churning out MIDI clocks to an Atari sequencer which then echoed them back to input E - which had no adverse effect, as the system continued to function.
A synchroniser converts either tape sync or SMPTE to MIDI song position pointer and timing information. Most have a built-in merge which can be used to good effect here. Simply connect the Master keyboard to the sync box, which will combine the note data with that internally generated for timing and send this to the sequencer. Alternatively, should the unit not have a merge facility, the one within the XR400 can be used instead. Connect its output to input A, B or C and the timing data will be merged with any information generated by the Master keyboard.
SYSTEM EXCLUSIVE information is one area of MIDI which is always likely to be fraught with problems. SysEx is the method by which MIDI devices transmit and receive their internal parameters, be that a single parameter change or an entire bulk dump of the voice parameters. There are practically as many protocols as there are manufacturers and the speed with which the MIDI device hardware can react and receive data is highly variable. Computer software written specifically for this often has to insert pauses to allow for a synth to "digest" the data received.
Hybrid Arts' Genpatch (a generic librarian) found itself sending and receiving data through the XR400 for the purposes of this review. Of the devices I tried, there were no problems with the Akai S900 and S950, Kawai K1, Roland D110, Yamaha TX7 and Casio CZ101. However, a Matrix 1000 had to be reset, because although the data was transmitted OK, it was corrupted when sent back. Similar problems occurred with a Roland MKS70 (JX10 module) and with a one-way dump from a Roland D50 (although the latter worked fine in handshake mode). The Matrix 1000 and the MKS70 have difficulty coping with SysEx MIDI data running at full tilt and it is quite possible that the XR400 would need fine-tuning to handle it and them. XRI say they're quite happy to look into any incompatibility problems.
Using more than one of a particular MIDI device presents no problems to the XR400 when each machine is given a separate ID. Roland allow each synth to have a "unit number" to distinguish it from another of the same model, but this isn't always the case. A dump request via the XR400 to a sampler will initiate the dump procedure on any other with its MIDI In and Out connected to the computer.
Another problem with SysEx is one which is beyond XRI's control. SysEx information always starts with the hexadecimal byte F0 followed by the Manufacturer's Identification code, data, and an F7 to finish off. The data can be anywhere from 5 or 6 bytes up to 64,000 and beyond. SysEx protocol does not allow for any MIDI note, performance or timing data to butt in between the F0 and F7. If this should happen, the SysEx transfer will be aborted. I tried injecting MIDI clock and notes at the same time as a SysEx dump and found that the results were variable. The results appeared to depend on which editor was being used to transmit the voice parameters back to the sound module. This situation is unlikely to occur in general practice as few people will want to send full memory dumps at the same time as playing a sequencer. You might want to send the odd edit buffer, but we're only talking about 150 to 500 bytes which didn't cause any trouble when tried.
WHAT WE ARE talking about here is a device which is, at present, unique. It is aimed at the working musician as opposed to a technophile who wants to push everything to its limits. As such, it works admirably. There need be no more MIDI lead mix-ups or patchbay programming problems each time you want to spend a couple of hours updating the libraries on your synths or samplers with a visual editor. No more swearing as you try to unravel the bird's nest of MIDI cables.
On the negative side, I would have liked to have seen more MIDI inputs on the unit. There are a couple of spaces on the rear of the unit and I understand that XRI are considering enlarging the choice outside the Master keyboard and sequencer to one from five instead of the present three. Apart from that, the XR400 is difficult to fault. It comes complete with an excellent manual which includes 14 diagrams of practical situations, along with a section on trouble-shooting and hints.
At an RRP of £200, this is likely to sell well. Perhaps the time is ripe for you to stop doing large amounts of damage to your MIDI leads and your sanity.
Price £199.95 including VAT
Review by Vic Lennard
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!