A huge thank you to Shueh-li of Xenovibes for being our first artist guest blogger, and for starting off with such an interesting and thoughtful post on Midi. There’s really no substitute for the perspective of an artist who has actually used technology to make a living. When your music and career’s on the line, that’s when you find out if a gizmo works or not.

However, I think it’s worth exploring some of the downsides of Midi in more detail.

It’s important to distinguish between Midi as a control protocol (eg, Midi pedals and mixers that use Midi to control parameters like volume and panning) and Midi as a medium for recording music performances. The former relies primarily on Midi continuous controller (CC) signals – turn a knob and a stream of info gets sent to gradually adjust a setting; while the latter relies on Midi note signals – press a key on a Midi keyboard and the corresponding note gets played by a synthesizer or sampler and/or recorded.

Nobody really has a problem with Midi controllers. It’s the latter use that’s fallen out of fashion somewhat. If you use a Midi keyboard to play a piano piece and record the resulting sound, there’s no issue. But if you only record the Midi data, that brings up some challenges.

At this point, I’ve got to emphasize that Midi isn’t sound; it’s instructions to play sound. It’s the equivalent of using a computer programming language like Java to tell your computer to play a song. When you record the Midi portion of a musical performance, you’re only recording the Midi instructions, not the resulting audio. That’s what makes Midi files so small and portable. Instructions take up a fraction of the space audio information would. Think of Midi as the high-tech equivalent of a piano roll – you know, those perforated drums mounted in pianos in those old cowboy movies that would rotate and auto-magically play back a piano piece.

As Shueh-li pointed out, Midi transmits information serially and at a particular speed (31.250 bits per second). That might sound like a lot of bandwidth, but when you factor in that for a single note played – say a high-C – the note itself, the volume, and modifiers such as velocity and aftertouch each count as individual instructions, Midi’s bandwidth no longer seems that copious. When you’re talking about elaborate arrangements of symphonic scope, that’s a very real limitation. Needless to say, my own bedroom experiments come nowhere close to exhausting Midi’s potential, and if it’s good enough for a professional musician like Shueh-li, it shows that if you understand your needs, Midi’s just fine.

Another reason Midi has a bad name is that Midi files played on early computers sounded god-awful. That’s because the Midi information was used to activate the most primitive of software music synthesizers, resulting in primitive bleeps that often didn’t particularly pleasant. This wasn’t a limitation inherent in Midi, however, but in those old synthesizer programs. Any of those early Midi files would sound immeasurably better when played back on today’s synthesizers. If a Midi file sounds bad, blame the synthesizer, not the Midi file. If a Midi file plays the wrong notes, then you can complain.

The last reason for Midi’s notoriety we’ll look at is it’s abuse by musicians. Like any other tool, it can be used to make music or ruin it. Midi has the ability to quantize notes – to re-order them to the nearest adjacent unit of timing. Let’s say you’ve played a drum beat out of time. Click quantize and it’s magically snapped to perfect rhythm. When abused, quantization can make music sound terribly robotic.

Midi’s reputation suffered at the hands of artist like BT (Brian Transeau), who dismissed Midi as antiquated and opted to work with pure audio, brandishing tools like loop sequencer Ableton Live, which at that point was audio-only. But Midi refused to die, and ironically enough was eventually incorporated into Ableton Live in an acknowledgement that Midi has an extremely useful place in any musician’s arsenal.

The first time I used a Midi sequencer to perform on stage, it was Propellerhead Software’s first version of Reason, which is a virtual instrument rack mated to a Midi-only sequencer. Midi’s efficiency allowed me to run my set on a low-powered Apple G3 iBook without a single glitch. Although I used audio-Midi sequencer Digital Performer in my home studio, there’s no way it would have run reliably on stage with my hardware, so I was very grateful for Midi-only arrangements. It’s an indication of Midi’s longevity that Reason’s sequencer continues to be Midi-only today, despite having progressed to version 4. Sure, people regularly mate Reason to an audio sequencer, but the roots of Reason – easily one of the most approachable and popular pieces of music software – remain squarely rooted in Midi.

That’s not to say implementations of Midi haven’t progressed over the years. The protocol itself has remained unchanged for an eternity, but whereas in the old days you needed physical Midi cables and bulky interfaces to get anything done, today Midi signals are usually routed through computer software and USB cables. And while my first guitar synthesizer rig had a Midi pickup tacked onto my guitar and fed into a rack interface that sprouted into a ridiculous jumble of wires, all of that can now be done with a single USB cable. Progress is a gratifying thing.