Plug a keyboard into a computer, press a key, and a piano sound comes out. It’s natural to assume the keyboard sent that sound down the cable. It didn’t. The cable carried no sound at all.
That’s the part of MIDI that confuses almost everyone, and once you see it, the rest falls into place.
It looks like a file format, but it isn’t
MIDI sounds like MP3 or WAV, a type of audio file. It’s not. MIDI doesn’t contain any sound. It contains instructions.
Think of the player pianos your great-grandparents might have had. They played music by themselves, fed by a spinning roll of paper with holes punched in it. The paper held no music, only directions: as the roll turned, the holes told the piano which keys to press, when, and how hard. The piano made all the sound. The paper just told it what to do.
MIDI is that paper roll, made digital. The name stands for Musical Instrument Digital Interface, and it’s a shared language for describing a performance.
When you press middle C on a MIDI keyboard, it sends a tiny message that says, roughly, “This note was pressed, and it was pressed medium-hard.” Later, another message says “That note was released.” That’s it. Whatever is on the other end, a piece of software or a synthesizer, reads the instructions and makes the actual sound.
Why that distinction matters
A recording of a singer or an acoustic guitar is audio: a captured sound wave, baked in like a photograph. Because it’s a finished picture of the sound, changing it afterward is hard. If the singer hits a flat note, you can’t just grab it and drag it up. Specialized tools can fix it, but you’re reworking a recording, and push too far and it starts to sound strange.
MIDI is closer to the sheet music than the recording. Because it’s only instructions, you can change almost anything after the fact:
- Swap the instrument. Play a melody on a piano, click once, and it becomes a string section, a saxophone, or a futuristic synthesizer. The notes stay the same; only the thing playing them changes.
- Fix a wrong note. Move one note without touching anything else.
- Change the timing. Lengthen a pause, or speed up the whole song without it turning into chipmunks, since there’s no recorded sound to stretch.
The player piano comparison holds here too. You could swap in a better piano and feed it the same paper roll. The performance is unchanged; the instrument is new.
Where it came from
MIDI arrived in the early 1980s to solve a practical problem. Synthesizers from different manufacturers couldn’t talk to each other, so a musician with keyboards from two companies was stuck. The industry agreed on one shared language, and it has barely changed since. That’s remarkable in technology, where most things are obsolete in five years.
Where you’ll find it
Mostly out of sight. MIDI sits underneath modern music production, film scoring, and video game audio. A USB keyboard plugged into a laptop, drum beats tapped out on a tablet, software controlling other software: all of it is a silent messenger carrying a human gesture to a computer that knows what to do with it.
So, to sum up
MIDI is a language for describing a musical performance, not a recording of one. It says which notes, when, and how hard, and leaves the sound to whatever is listening. That’s why it’s so flexible: you can change the instrument, the timing, or a single wrong note without re-recording a thing.
It’s the paper roll, not the piano. And as the player piano showed a century ago, the roll is the part you can edit.