The MIDI specification first saw the light of day at the 1981 AES, when Dave Smith of Sequential Circuits presented a paper on the “Universal Synthesizer Interface.” It was co-developed with other companies (an effort driven principally by Roland’s Ikutaro Kakehashi, a true visionary of this industry), and made its prime time debut at the 1983 Los Angeles NAMM show, where a Sequential Circuits Prophet-600 talked to a Roland keyboard over a small, 5-pin cable. I saw Dave Smith walking around the show and asked him about it. “It worked!” he said, clearly elated—but I think I detected some surprise in there as well.

“It” was the Musical Instrument Digital Interface, known as MIDI. Back in those days, polyphonic synthesizers cost thousands of dollars (and “polyphonic” meant 8 voices, if you were lucky and of course, wealthy). The hot computer was a Commodore-64, with a whopping 64 kilobytes of memory—unheard of in a consumer machine (although a few years before, an upstart recording engineer named Roger Nichols was stuffing 1MB memory boards in a CompuPro S-100 computer to sample drum sounds). The cute little Macintosh hadn’t made its debut, and as impossible as it may seem today, the PC was a second-class citizen, licking its wounds after the disastrous introduction of IBM’s PCjr.

Tom Oberheim had introduced his brilliant System, which allowed a drum machine, sequencer, and synthesizer to talk together over a fast parallel bus. Tom feared that MIDI would be too slow. And I remember talking about MIDI at a Chinese restaurant with Dave Rossum of E-mu systems, who said “Why not just use Ethernet? It’s fast, it exists, and it’s only about $10 to implement.”

But Dave Smith had something else in mind: An interface so simple, inexpensive, and foolproof to implement that no manufacturer could refuse. Its virtues would be low cost, adequate performance, and ubiquity in not just the pro market, but the consumer one as well.

Bingo.

But it didn’t look like success was assured at the time; MIDI was derided by many pros who felt it was too slow, too limited, and just a passing fancy. 30 years later, though, MIDI has gone far beyond what anyone had envisioned, particularly with respect to the studio. No one foresaw MIDI being part of just about every computer (e.g., the General MIDI instrument sets). This trend actually originated on the Atari ST—the first computer with built-in MIDI ports as a standard item (see “Background: When Amy Met MIDI” toward the end of this article).

Evolution of a spec

Oddly, the MIDI spec officially remains at version 1.0, despite significant enhancements over the years: the Standard MIDI File format, MIDI Show Control (which runs the lights and other effects at Broadway shows like Miss Saigon and Tommy), MIDI Time Code to allow MIDI data to be time-stamped with SMPTE timing information, MIDI Machine Control for integration with studio gear, microtonal tuning standards, and a lot more. And the activity continues, as issues arise such as how best to transfer MIDI over USB, with smart phones, and over wireless.

The MIDI Manufacturers Association

The guardian of the spec, the MIDI Manufacturers Association (MMA), has stayed a steady course over the past several decades, holding together a coalition of mostly competing manufacturers with a degree of success that most organizations would find impossible to pull off. The early days of MIDI were a miracle: in an industry where trade secrets are jealously guarded, manufacturers who were intense rivals came together because they realized that if MIDI was successful, it would drive the industry to greater success. And they were right. The MMA has also helped educate users about MIDI, through books and online materials such as “An Introduction to MIDI.”

I had an assignment at the time from a computer magazine to write a story about MIDI. After turning it in, I received a call from the editor. He said the article was okay, but it seemed awfully partial to MIDI, and was unfair because it didn’t give equal time to competing protocols. I tried to explain that there were no competing protocols; even companies that had other systems, like Oberheim and Roland, dropped them in favor of MIDI. The poor editor had a really hard time wrapping his head around the concept of an entire industry willingly adopting a single specification. “But surely there must be alternatives.” All I could do was keep replying, “No, MIDI is it.” Even when we got off the phone, I’m convinced he was sure I was holding back information on MIDI’s competition.

MIDI HERE, MIDI THERE, MIDI EVERYWHERE

Now MIDI is everywhere. It’s on the least expensive home keyboards, and the most sophisticated studio gear. It’s a part of signal processors, guitars, keyboards, lighting rigs, smoke machines, audio interfaces…you name it. It has gone way beyond its original idea of allowing a separation of controller and sound generator, so people didn’t have to buy a keyboard every time they wanted a different sound.

SO WHERE’S IT GOING?

“Always in motion, the future…” Well, Yoda does have a point. But the key point about MIDI is that it’s a hardware/software protocol, not just one or the other. Already, the two occasionally take separate vacations. The MIDI data in your DAW that drives a soft synth doesn’t go through an opto-isolators or cables, but flies around inside your computer.

One reason why MIDI has lasted so long is because it’s a language that expresses musical parameters, and these haven’t changed much in several centuries. Notes are still notes, tempo is still tempo, and music continues to have dynamics. Songs start and end, and instruments use vibrato. As long as music is made the way it’s being made, the MIDI “language” will remain relevant, regardless of the “container” used to carry that data. However, MIDI is not resting on its laurels, and neither is the MMA—you can find out what they’re working on for the future here.

Background: When Amy Met MIDI