The ‘Scores of Scores’ lieder corpus encoding project is a major contribution to the OpenScore encoding effort. We are creating and releasing publicly a large repository of high-quality musical scores which any internet user will be able to download and use as they please.

  1. This page (an introduction).
  2. The Notes for transcribers
  3. ‘Learn by Doing Lieder Transcription’: Exercises for Intrepid Students
  4. Homepage for the launch event, ‘Scores of Scores’: Possibilities and Pitfalls with Musical Corpora.

Overview and motivation

IMSLP has led the way with an initiative to provide the world with a huge repository of over 125,000 out-of-copyright musical scores in PDF (image) format. This has greatly democratised access to music, and enabled musicians to explore the repertoire more widely, more easily. This is great, but people are able to engage with music in much more direct and varied ways when they have access to ‘encoded’ score: not PDFs, but files which can be opened in free notation software like ‘MuseScore’ and adapted for specific tasks. The difference is the same as that between looking at an image of text and being able to edit that text in a word-processor. There are some corpora of scores in these editable formats, but they are inconsistent in their coverage and quality. Similarly, the existing ‘Optical Music Recognition’ software for converting from image to encoded format is not yet up to the task at scale and will be helped along by having reliable image-encoding pairs to work with.

The many popular uses of encoded scores include extracting a single part for private practice, creating arrangements for whatever musical forces are available, and making bespoke teaching resources (an ongoing focus for ‘Four Score and More’). It is also be useful to ‘Music Information Retrieval’: an academic field that analyses structural patterns in music.

Collaborators and engagement

We aim to engage with the widest possible public: anyone interested in music, of any age, ability, and nationality. To that end of engaging with that wide audience, we are delighted to have arranged this collaboration with OpenScore and IMSLP (the Petrucci Music Library). OpenScore have connected us with their awesome community of encoders, and both sites will host the final scores.

The online repository ensures that there is no geographical barrier to encoding or accessing the scores. Just as importantly, there is also no financial barrier to entry thanks to MuseScore: a free, open-source, and high-quality notation software which can import/export to musicXML, and by extension to all industry-standard file types. Perhaps most remarkably of all, even blind musicians will be able to access the scores through the music21-enabled Braille output.

How can people get involved?

Everyone is welcome to contribute to the encoding effort through OpenScore. Here’s the information for prospective contributors: OpenScore: How You Can Get Involved.

Apart from helping the musical world, encoders themselves also benefit from the task: copy-editing music is a very worthwhile pursuit that musicians have used as part of their education for as long as records exist. The story of Wagner copying out Beethoven’s Ninth Symphony twice stands as a famous testament to this. So, encode on and perhaps someday you’ll write world-changing, epic operas!

The Lieder Corpus project is an effort to transcribe hundreds of German and French songs, and to make them available copyright-free (under a CC0 license)

Launch event

In association with the Cambridge Big Data Strategic Research Initiative, Cambridge Digital Humanities Network and the University of Cambridge Faculty of Music, a ‘Scores of Scores’ event took take place on 19 June 2018, combining elements of a conference, workshop, hackathon, and ‘encodathon’. This saw the launch of both the ‘Scores of Scores’ encodings, and also a first application of those scores: the ‘Cut Outs’ automatic music theory exercises.