What is music21?

Music21 is a Python-based toolkit for computer-aided musicology.

People use music21 to answer questions from musicology using computers, to study large datasets of music, to generate musical examples, to teach fundamentals of music theory, to edit musical notation, study music and the brain, and to compose music (both algorithmically and directly).

One of music21’s mottos is “Listen Faster.” With the toolkit you should be able to find interesting moments and get a sense of the overall profile of a piece or a repertory of pieces. We hope that with the computer you’ll have more time for listening and playing for enjoyment and use less of your time listening for work.

The system has been around since 2008 and is constantly growing and expanding. The approaches and traditions in music21 have been used in many previous software systems. See Authors, Acknowledgments, Contributing, and Licensing for information on the authors and background of the project.

The 21 in music21 refers to its origins as a project nurtured at MIT. At MIT all courses have numbers and music, along with some other humanities departments, are numbered 21. The music departments of MIT, along with Harvard, Smith, and Mount Holyoke Colleges, helped bring this toolkit from its easiest roots to a mature system.

Finding solutions in a hurry

With music21 adds a collection of specialized tools and objects to the general-purpose and easy to understand “Python” programming language. Install music21 and type python3 (or, better, ipython) and load it by typing:

from music21 import *

…and thousands of musical tools become available to you. For instance, want to see a note on the screen? Type these lines:

n = note.Note("D#3")
n.duration.type = 'half'
n.show()
../_images/what_6_0.png

Need a whole line of notes? Even easier:

littleMelody = converter.parse("tinynotation: 3/4 c4 d8 f g16 a g f#")
littleMelody.show()
../_images/what_8_0.png

Want to hear the melody? It’s just as easy! (Please give it a second or two after hitting play for the piano sounds to load):

littleMelody.show('midi')

Want to view the opening tone-row of Schoenberg’s Fourth String quartet as a matrix?

print(serial.rowToMatrix([2, 1, 9, 10, 5, 3, 4, 0, 8, 7, 6, 11]) )

   0 11  7  8  3  1  2 10  6  5  4  9
   1  0  8  9  4  2  3 11  7  6  5 10
   5  4  0  1  8  6  7  3 11 10  9  2
   4  3 11  0  7  5  6  2 10  9  8  1
   9  8  4  5  0 10 11  7  3  2  1  6
  11 10  6  7  2  0  1  9  5  4  3  8
  10  9  5  6  1 11  0  8  4  3  2  7
   2  1  9 10  5  3  4  0  8  7  6 11
   6  5  1  2  9  7  8  4  0 11 10  3
   7  6  2  3 10  8  9  5  1  0 11  4
   8  7  3  4 11  9 10  6  2  1  0  5
   3  2 10 11  6  4  5  1  9  8  7  0

Get a quick graph showing how common various pitches are in a fourteenth century piece:

dicant = corpus.parse('trecento/Fava_Dicant_nunc_iudei')
dicant.plot('histogram', 'pitch')

dicant.measures(1, 10).show()
../_images/what_15_0.png

This example, and many below, come from the music21 built-in corpus of thousands of pieces that come with the system to help you get started right from the beginning. We believe in “Batteries Included” as a core principle. So for instance, every Bach chorale is included, so that you can do things like add the note name in german to every note in Bach chorale, BWV295:

bwv295 = corpus.parse('bach/bwv295')
for thisNote in bwv295.recurse().notes:
  thisNote.addLyric(thisNote.pitch.german)
bwv295.show()
../_images/what_17_0.png

Prepare an incipit index (thematic catalog) of every Bach chorale that is in 3/4: (we’ll just look at the first 25 here)

catalog = stream.Opus()
for work in corpus.chorales.Iterator(1, 26):
    firstTimeSignature = work.parts[0].measure(1).getTimeSignatures()[0]
    if firstTimeSignature.ratioString == '3/4':
        incipit = work.measures(0,2)
        catalog.insert(0, incipit.implode())

catalog.show()
../_images/what_19_0.png ../_images/what_19_1.png ../_images/what_19_2.png ../_images/what_19_3.png

Advanced analysis tools are included. Want to know how unstable the rhythmic profile of a piece is? Use Ani Patel’s nPVI function on it:

s = corpus.parse('AlhambraReel')
analysis.patel.nPVI(s.flatten())

 5.755395683453239

Learning music21

Music21 can be simple to use but it is also extremely powerful. Like all powerful software (Photoshop compared to MS Paint, AutoCAD, Excel), there’s a bit of a learning curve, especially for people who haven’t programmed before.

To use music21, some familiarity with the “Python” programming language is needed. Python is widely regarded as one of the easiest languages to learn and is often taught as a first programming language. You don’t need to be a seasoned programmer; just a little bit of Python and you will be able to get started and explore music in new ways with music21.

Probably the hardest thing about music21 is getting it installed and writing the first line of code. The installation instructions at Installing music21 will help you get started, and then we can continue with the rest of the User’s Guide.

If you need help at any time, there are always helpful music21 fanatics at the mailing list, https://groups.google.com/g/music21list/.

Continue on to Installing music21 or learn more about who made the system and who supported it.