This is a README file for a data repository originating from the DCML corpus initiative and serves as welcome page for both

• the GitHub repo https://github.com/DCMLab/ABC and the corresponding
• documentation page https://dcmlab.github.io/ABC

For information on how to obtain and use the dataset, please refer to this documentation page.

When you use (parts of) this dataset in your work, please read and cite the accompanying data report:

Neuwirth, M., Harasim, D., Moss, F. C., & Rohrmeier, M. (2018). The Annotated Beethoven Corpus (ABC): A Dataset of Harmonic Analyses of All Beethoven String Quartets. Frontiers in Digital Humanities, 5(July), 1–5. https://doi.org/10.3389/fdigh.2018.00016

The ABC forms part of the larger Distant Listening Corpus which constitutes a data infrastructure the data report of which has implications for the present corpus, too:

Hentschel, J., Rammos, Y., Neuwirth, M., & Rohrmeier, M. (2025). A corpus and a modular infrastructure for the empirical study of (an)notated music. Scientific Data, 12(1), 685. https://doi.org/10.1038/s41597-025-04976-z

The Annotated Beethoven Corpus (ABC) (A corpus of annotated scores)

The ABC dataset consists of expert harmonic analyses of all Beethoven string quartets (opp. 18, 59, 74, 95, 127, 130, 131, 132, 135, composed between 1800 and 1826), encoded in a human- and machine-readable format (MuseScore format). Using a modified Roman Numeral notation (the DCML harmony annotation standard), the dataset includes the common music-theoretical set of harmonic features such as key, chordal root, chord inversion, chord extensions, suspensions, and others.

Getting the data

• download repository as a ZIP file
• download a Frictionless Datapackage that includes concatenations of the TSV files in the four folders (measures, notes, chords, and harmonies) and a JSON descriptor:
  • ABC.zip
  • ABC.datapackage.json
• clone the repo: git clone https://github.com/DCMLab/ABC.git

Data Formats

Each piece in this corpus is represented by five files with identical name prefixes, each in its own folder. For example, the first movement of the first quartet, op. 18/1, has the following files:

• MS3/n01op18-1_01.mscx: Uncompressed MuseScore 3.6.2 file including the music and annotation labels.
• notes/n01op18-1_01.notes.tsv: A table of all note heads contained in the score and their relevant features (not each of them represents an onset, some are tied together)
• measures/n01op18-1_01.measures.tsv: A table with relevant information about the measures in the score.
• chords/n01op18-1_01.chords.tsv: A table containing layer-wise unique onset positions with the musical markup (such as dynamics, articulation, lyrics, figured bass, etc.).
• harmonies/n01op18-1_01.harmonies.tsv: A table of the included harmony labels (including cadences and phrases) with their positions in the score.

Each TSV file comes with its own JSON descriptor that describes the meanings and datatypes of the columns ("fields") it contains, follows the Frictionless specification, and can be used to validate and correctly load the described file.

Opening Scores

After navigating to your local copy, you can open the scores in the folder MS3 with the free and open source score editor MuseScore. Please note that the scores have been edited, annotated and tested with MuseScore 3.6.2. MuseScore 4 has since been released which renders them correctly but cannot store them back in the same format.

Opening TSV files in a spreadsheet

Tab-separated value (TSV) files are like Comma-separated value (CSV) files and can be opened with most modern text editors. However, for correctly displaying the columns, you might want to use a spreadsheet or an addon for your favourite text editor. When you use a spreadsheet such as Excel, it might annoy you by interpreting fractions as dates. This can be circumvented by using Data --> From Text/CSV or the free alternative LibreOffice Calc. Other than that, TSV data can be loaded with every modern programming language.

Loading TSV files in Python

Since the TSV files contain null values, lists, fractions, and numbers that are to be treated as strings, you may want to use this code to load any TSV files related to this repository (provided you're doing it in Python). After a quick pip install -U ms3 (requires Python 3.10 or later) you'll be able to load any TSV like this:

import ms3

labels = ms3.load_tsv("harmonies/n01op18-1_01.harmonies.tsv")
notes = ms3.load_tsv("notes/n01op18-1_01.notes.tsv")

Version history

See the GitHub releases.

Questions, Suggestions, Corrections, Bug Reports

Please create an issue and/or feel free to fork and submit pull requests.

Publications

• The accompanying Data Report has been published in Neuwirth, M., Harasim, D., Moss, F. & Rohrmeier M. (2018).
• An evaluation of the dataset can be found in this Moss, F., Neuwirth M., Harasim, D. & Rohrmeier, M. (2019).
• The latest version of the annotation standard has been described in Hentschel, J., Neuwirth, M. & Rohrmeier, M. (2021)

Cite as

Neuwirth, M., Harasim, D., Moss, F. C., & Rohrmeier, M. (2018). The Annotated Beethoven Corpus (ABC): A Dataset of Harmonic Analyses of All Beethoven String Quartets. Frontiers in Digital Humanities, 5(July), 1–5. https://doi.org/10.3389/fdigh.2018.00016

License

Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0).

Remarks

Hybrid version of annotations

v1.0 of this dataset used version 1.0.0 of the DCML harmony annotation standard, with the version number attributed only in hindsight. Some of the labels that were corrected in the meantime, however, use features available only in later versions, such as recursive reference to a secondary ('tertiary') key as in V7/V/V. However, changes have been applied conservatively in order to keep the data as homogeneous as possible.

Possible annotation errors

While the annotation process (as detailed in the Data Report) was conducted very carefully, and all annotated symbols have been automatically checked for syntactic correctness, the authors cannot entirely rule out the possibility of typograpical errors or misinterpretations (e.g., due to ambiguities in the score or underspecification in the musical texture). After all, music analysis is not a deterministic process but involves interpretation. If you encounter anything that seems not right to you, please create an issue here.

Missing bars

The original XML file for Op. 132 No. 15, mov. 5 from Project Gutenberg did not contain measures 194-241. We added them manually.

File naming convention

n(?P<quartet>\d{2})  # quartet number, e.g. n01
op(?P<op>\d{2,3})    # opus number, e.g. op18
(?:-(?P<no>\d))?     # (optional) number within the opus, e.g. -1
_(?P<mvt>\d{2})      # movement number, e.g. _01

Overview

| file_name |measures|labels|standard| annotators |reviewers | |------------|-------:|-----:|--------|---------------|----------| |n01op18-1_01| 313| 405|1.0.0 |Markus Neuwirth| | |n01op18-1_02| 110| 263|1.0.0 |Markus Neuwirth| | |n01op18-1_03| 145| 203|1.0.0 |Markus Neuwirth| | |n01op18-1_04| 381| 598|1.0.0 |Markus Neuwirth| | |n02op18-2_01| 249| 486|1.0.0 |Markus Neuwirth| | |n02op18-2_02| 86| 177|1.0.0 |Markus Neuwirth| | |n02op18-2_03| 87| 138|1.0.0 |Markus Neuwirth| | |n02op18-2_04| 412| 468|1.0.0 |Markus Neuwirth| | |n03op18-3_01| 269| 383|1.0.0 |Markus Neuwirth| | |n03op18-3_02| 151| 394|1.0.0 |Markus Neuwirth| | |n03op18-3_03| 168| 278|1.0.0 |Markus Neuwirth| | |n03op18-3_04| 364| 569|1.0.0 |Markus Neuwirth| | |n04op18-4_01| 219| 554|1.0.0 |Markus Neuwirth| | |n04op18-4_02| 261| 369|1.0.0 |Markus Neuwirth| | |n04op18-4_03| 98| 145|1.0.0 |Markus Neuwirth| | |n04op18-4_04| 217| 386|1.0.0 |Markus Neuwirth| | |n05op18-5_01| 225| 430|1.0.0 |Markus Neuwirth| | |n05op18-5_02| 105| 168|1.0.0 |Markus Neuwirth| | |n05op18-5_03| 139| 247|1.0.0 |Markus Neuwirth| | |n05op18-5_04| 300| 567|1.0.0 |Markus Neuwirth| | |n06op18-6_01| 264| 374|1.0.0 |Markus Neuwirth| | |n06op18-6_02| 79| 265|1.0.0