MNX 1.0 draft specification

MNX is a new, open standard for representing music notation for interchange and internal use in software applications. It builds on the ideas and success of MusicXML (and is maintained by the same people) but aims to avoid the design choices of MusicXML that prevent software from using it as a native format or to present the same underlying musical semantics in different forms (such as reflowing score or part layouts). And MNX represents everything in JSON, the easy-to-use, dominant standard for information exchange among Internet applications.

MNX is still a work in progress without a stable standard for implementation. But these pages will show you what we’ve developed so far. We encourage you to give feedback via our GitHub discussions forum.

Why another music notation format?

If this is your first introduction to the concept of representing music with a computer, we encourage you to read the MusicXML tutorial, which lays out the motivations and key ideas of music representation in the most widely used symbolic notation format, and one that you can use today. MNX is different from MusicXML (or abc, MEI, MIDI, or any number of other good formats), but the same basic ideas and problems arise in all music representations, so knowing about one of them before reading about MNX is a good place to start.

You might ask: if there are already dozens, if not hundreds, of music representation formats, why create another one?

Most, if not all, existing notation formats were designed first to store music in a way that it could be exchanged with others — other people or other software programs. And they were designed to represent the look of music on a page, once it’s printed.

MNX will do that, of course, but it differs in that it’s first and foremost designed to store and exchange semantic elements of music — the notes, rhythms, pitches, parts — and to allow multiple layouts to emerge or be defined dynamically. Beautiful scores on paper are always important, but these days music notation is displayed and consumed in many other ways. A piece might be viewed in a continuous scroll, or a musician might view her part on a tablet of different dimensions than the computer it was engraved on. A blind or visually-impaired user might require different magnifications or Braille. A modern format needs to make these types of workflows not just possible, but easy.

MNX is also designed to be easy for software to read, write and pass through. It uses JSON, the most widely used format for data exchange on the web, which maps easily to object models not just in JavaScript but every modern programming language. Its objects map closely to how concepts tend to be implemented in applications. And our philosophy is to provide clear, unambiguous ways to encode musical concepts — “one and only one way to do things,” hence helping notation applications avoid bugs and making interchange more reliable.

A gentle introduction

Conceptually, MNX is a way to represent this image...

...as something like this:

staves=1
keysignature=C major
clef=treble
notes=whole note at C4

The former, an image, is intended for humans. A computer can’t understand it without using advanced, error-prone computer vision techniques.

The latter — which encodes the meaning (aka semantics) of the music, as opposed to a specific collection of pixels — is much easier for a computer to understand. Treating music in this way means it can be shared among different programs, automatically manipulated and much more.

MNX is a standard that describes how to encode music, so that computers can have a shared understanding on how to interpret it. As HTML describes a web page such that any web browser can read it, MNX describes a piece of music such that any notation software can read or write it.

Motivation

People have been using computers to encode music for decades; a wide variety of encoding formats has been developed over the years. Why do we need yet another format?

Each music encoding format has its own biases and priorities. The ABC format, for example, aims for simplicity and was originally developed for folk music. MusicXML is optimized for notation interchange. MIDI focuses on low-level instrument sequencing, as opposed to higher-level musical concepts. MEI, a general-purpose framework for encoding arbitrary musical documents, pays particular attention to the needs of scholars. Proprietary formats, such as Finale’s `.musx` or Sibelius’ `.sib`, are optimized for specific notation programs and not meant to be read by others.

Our goal with MNX is to create a format that does all of the following:

The existing format that comes closest to these goals is MusicXML. In fact, MNX’s creation is being led by the W3C Music Notation Community Group, which also oversees the MusicXML standard. So what’s the difference?

We see MNX as the next generation of MusicXML, enabling new uses that MusicXML didn’t set out to support.

With hundreds of applications supporting it, MusicXML has succeeded in becoming the de facto notation interchange standard. These days, you can be reasonably sure that, if you compose music in a commonly used notation application, you’ll be able to export it and open it in another application with decent fidelity — thanks to MusicXML. This has been a major cultural step forward, considering major notation programs hadn’t always looked kindly on making music exportable into competitors’ products.

As the dream of universal interchange has become a reality, users and developers have come to demand even more. Users want richer interchange (i.e., preserving the details of music more crisply), the advent of “reflowing” web-based renderers has shifted expectations around music engraving, and some developers even use MusicXML as a native format (something it wasn’t designed to do).

MNX is an effort to take everything we’ve learned from 15 years of MusicXML development — along with its diverse community of developers, musicians and publishers — and build a better future together.

Finally, we should note there are many music notation systems throughout the world’s cultures, along with historical notation systems. MNX deals specifically with Common Western Musical Notation, as it’s a high priority for the current MusicXML community and it’s where our collective expertise is the strongest — but our long-term plan is to support other systems. We envision adding other formats to the MNX family, though we’re not sure what they’ll be or how they’ll be named.

Getting started

Curious about what we have so far? The best way to dive in is to read Comparing MNX and MusicXML. Even if you're not familiar with MusicXML, the document has many examples of sheet music with the corresponding MNX code.

Want to contribute to our work? MNX is being designed in an open process that you're encouraged to join. Here's how:

Copyright © 2021 the Contributors to the MNX Specification, published by the Music Notation Community Group under the W3C Community Contributor License Agreement (CLA). A human-readable summary is available.