What is MIDI? That's how we're kicking off a beginning tutorial series on Synth Expert? Not how to create a feedback loop in your modular synth that causes your studio to time travel? Or at least, sound like the TARDIS? Not how to program sequences from Stranger Things? That's right, and here's why: When I talk to tech support personnel, product clinicians, and synth makers about what sorts of how-to articles they'd like to see on this site, their first response is almost always, "Clear up the confusion about MIDI!"
With MIDI (Musical Instrument Digital Interface) having been with us for just about 35 years now, and being the backbone of how most musical instruments (if they use electricity at all) talk to each other and even internally to themselves, you'd think there wouldn't be much left to be confused about. But call centers are routinely beset with inquiries about, for example, why a MIDI file sounds different when played back on different instruments, why a MIDI track in a DAW project isn't making any sound, and even (this actually happened) why a user couldn't play back a MIDI project from the USB port for his car stereo.
By way of an oversimplified and brief origin story (insofar as I'm capable of being brief; I've been told otherwise), in the early '80s many synth makers had their own protocols and connectors for interconnecting their gear. But while, say, a Roland synth could communicate with a Roland sequencer or an Oberheim with an Oberheim, different brands of gear couldn't talk to each other. In a rare and still unduplicated act of cooperation, a bunch of folks in the industry reasoned that universal compatibility would help all manufacturers' sales. For example, if someone already had brand X of synth but their local music store only carried brand Y of sequencer, they might forego buying any sequencer because of not wanting to mail-order brand X sight unseen. (Keep in mind these were days when there was no internet on which to research gear.)
It's well known that Dave Smith of Prophet-5 fame drew up most of the MIDI spec and Ikutaro Kakehashi threw the weight of Roland behind it all, then other companies felt they had to be onboard, and, as my British publishers would say, Bob's your uncle. The 1983 NAMM show featured the two first MIDI polysynths - a Prophet-600 and a Roland Jupiter-6 - communicating, and people's minds were blown when pressing a key on one synth actually played notes on both! (This setup was duplicated in 2013 for MIDI's 30th anniversary.) Though MIDI had nearly endless applications, for a long time people mainly associated it with having "extra hands" and being able to make fat stacks of sound using multiple synths.
Okay, back to business. That early NAMM example illustrates something: If you'd unplugged the audio line on either of those synths but left the MIDI cable in place, you would have stopped hearing whichever synth you unplugged, even though MIDI data was still traveling between the two. This brings us to the all-important first commandment of MIDI:
MIDI. Is. Not. Audio.
Unlike the bits in an MP3 file, the pits on a CD, the grooves on a record, or the voltage coming from the 1/4'' outputs of your synthesizer, MIDI carries no audio information. It is not, itself, music. Instead, it's a set of instructions that tell something that can create audio/music what to do. (It can also tell non-musical machines like stage lighting rigs what to do, but I'm getting ahead of myself.)
In DAWs, you'll often hear about "piano roll" MIDI editing. This term comes from vintage player pianos, which used rolls of paper punched with holes. Mechanical sensors would read where the holes were and trigger the right notes. In a way, the piano roll was like a primitive MIDI file. Point being, it needed a piano to execute its instructions; you wouldn't put one on a record player (or scan it with a laser) and expect to hear anything.
Likewise, MIDI data needs to be interpreted by something that understands it and puts out the audio accordingly. Fortunately for us, nearly everything understands MIDI, but the "portability" of MIDI is also the source of some of the confusion. You can record a MIDI file using the pristine sounds of, say, a recent Yamaha Motif, then put that MIDI file into an off-the shelf computer with a consumer-grade sound card. The right notes will play, after a fashion, but you're going to hear different instrument sounds because a different audio device is making them.
Similarly, if you import a MIDI track into your DAW, but that track is not linked to any virtual instrument or external hardware synth, you can hit the space bar and the play wiper will silently skip down the lane. (Most DAWs have a dedicated instrument track type that takes MIDI as input and gives you the soft synth's audio as output.) What's more, MIDI doesn't magically move audio between two connected instruments; one of my early adolescent synth mistakes was thinking it could and that I only needed to plug one of my two synths into the amp! (This may have been motivated by the amp only having one functioning input. My next mistake was thinking a Y-cable would make a passable mixer.)
I'm running long so I'll finish up this installment with a few things people might mean when they say MIDI.
- The MIDI spec itself: This is the totality of the set of instructions (notes on and off, continuous control messages, and more) that make up MIDI, the numerical rules manufacturers have to follow to make it all work. "Dave Smith and Kakehashi-san invented MIDI."
- MIDI data itself: "MIDI is going out of my Motif and into my Kurzweil." This could mean something as simple as you noodling on the keys with a MIDI cable connected, or it could involve a complex multitrack project you're driving from a sequencer MIDI data generated in real time, as from just playing a keyboard, isn't necessarily a MIDI file. Which brings us to ...
- A MIDI file: "Wow, this MIDI file is huge!" This is a bunch of MIDI data that's saved in a storage medium (internal sequencer memory, your hard drive, USB stick, etc.) and can be recalled for later use.
- USB: Here's a smaller but equally re-occurring confusion. Many controller keyboards and other devices connect to a computer directly via USB; you don't need a separate MIDI interface to use them. This leads some people to think that USB is something entirely different than the old-school 5-pin connectors. And while it does move the ones and zeros around a bit differently, the point is: MIDI over USB is still MIDI. It works the same and all the rules apply. Some USB connections might carry audio and MIDI at the same time over a single cable, but these are separate data streams so the MIDI is still not audio.
A final fun fact: The MIDI spec is still at version 1.0, and still works fine. How many things that are over 30 years old can that be said about? It's even seamless across gear from wildly different eras. The 2013 NAMM display I mentioned earlier also featured a Commodore 64 computer being sequenced by an iPad (or maybe it was the other way around) with no problems.
Next time, we'll talk about the different types of MIDI data. After that, we'll talk about why MIDI still matters even though nowadays even a midrange smartphone has enough CPU power to process many tracks of high-resolution audio. A huge thing that cemented MIDI's popularity is that this wasn't always the case: In the home studio of the 1990s, MIDI was the only way to affordably have a multi-track experience. You MIDI-sequenced all your synths and modules from a hardware sequencer or rudimentary software program, recorded all those audio outputs in real time, and mixed everything down to a stereo tape recorder or, if you were lucky, an Alesis ADAT.
More about this next time. If you'd like to dig in further in the meantime, browse around at The MIDI Association (www.midi.org). They're the non-profit organization that acts as official steward of the MIDI spec and engages in myriad industry outreach and public education efforts.