What is MIDI?

How MID Works

A .mid file is a Standard MIDI File, and the single most important fact about it is that it contains no recorded audio whatsoever. There is no waveform inside a MIDI file, no samples, nothing that captures the actual sound of an instrument. Instead, a MID file stores a sequence of performance events: instructions that describe what a musician did, not the sound that resulted. This is why MIDI is often described as sheet music for machines rather than a recording.

The events themselves are compact messages. The core ones are note-on and note-off, which say that a particular pitch started or stopped, together with a velocity value that typically represents how hard a key was struck. Alongside these, a MIDI stream carries program change messages that select which instrument sound should be used, and controller messages that adjust things like volume, sustain, or modulation. Every event is stamped with timing information so the sequence unfolds correctly over time.

Because a MID file stores instructions rather than sound, the actual audio is generated at playback time. A synthesizer or a software sound bank reads the note events and produces the tones, using its own set of instrument voices. The consequence is striking: the same MIDI file can sound completely different on two different devices, because each one interprets the instructions with its own sounds. A file that sounds rich on one synthesizer may sound thin or simply different on another.

This design is also why MIDI files are extraordinarily small. Storing a list of note events costs a tiny fraction of what it takes to store a sampled waveform, so an entire multi minute arrangement can fit in a handful of kilobytes. It also makes the music fully editable at the level of individual notes. A composer can change a single pitch, shift timing, swap the instrument assigned to a track, or transpose an entire part, all by editing the instructions rather than manipulating audio.

History and Standardization

MIDI, short for Musical Instrument Digital Interface, emerged in the early 1980s as a way for electronic instruments from different manufacturers to communicate. It was created through cooperation among synthesizer makers who agreed on a common protocol so that a keyboard from one company could control a sound module from another. This origin as a hardware interconnection standard is why MIDI is fundamentally about messages and events rather than audio.

The Standard MIDI File format extended that live protocol into a way of saving performances to disk, giving the .mid extension its meaning. It preserved the same event based approach, so a saved file is essentially a recording of the control messages rather than of the sound. This let sequences be exchanged between programs and devices, which was crucial in an era of many incompatible music systems.

One recurring problem was that program change messages did not guarantee the same instrument on every device. General MIDI addressed this by standardizing a common set of instrument assignments, so that a given program number would map to a broadly consistent sound across compliant devices. This improved the odds that a file would play back with roughly the intended instrumentation, although the exact tone still depended on each device's sound bank.

MID Compared to Recorded Audio Formats Like WAV and MP3

Comparing MID to WAV or MP3 is really a comparison between two different ideas of what an audio file is. WAV stores an uncompressed waveform and MP3 stores a compressed one, but both hold the actual sound. Play either on any device and, aside from minor hardware differences, it sounds the same everywhere. MIDI holds instructions instead, so its playback is only as good, and only as consistent, as the synthesizer interpreting it.

That difference drives every tradeoff. MIDI wins decisively on size and on editability: a complex piece can be a few kilobytes, and every note remains changeable after the fact. Recorded formats win decisively on fidelity consistency: what you hear is fixed and reproducible, capturing real instruments, vocals, and nuance that a synthesizer cannot recreate from a note list. A singer's voice, for instance, simply cannot be represented in MIDI at all.

The honest way to frame it is that these formats are not competitors so much as tools for different jobs. MIDI is ideal when you want a tiny, fully editable representation of a musical arrangement and can accept that its sound depends on the playback device. WAV and MP3 are the choice when you need a faithful, self contained recording that sounds identical everywhere. Choosing between them is really choosing whether you value flexibility and size or guaranteed fidelity.