Project report

Dynamic scoring-up and display of early music in Verovio

How do you read old scores? How do these historic pieces sound at all? Our research makes it easier to answer these questions!

Project participants

Anna Viktoria Katrin Plaksin
Martha Thomae PhD (Universidade NOVA de Lisboa)

PD Dr. Laurent Pugin (RISM Digital Center / Universität Bern

Project status

completed

We developed new features that make it possible to better visualize and hear early music, as written in the Middle Ages or Renaissance, in digital form. It is possible to dynamically switch between different views of musical pieces, for example, between individual voices and the full score. This not only makes the work of music researchers easier, as the automation of score creation allows them to concentrate more on philological and content-related work, but also gives students and musicians an easier access point to historical music. This makes early music easier to understand and to perform in its original form, helping to preserve it as cultural heritage.

The scoring-up feature for mensural notation is available in the new 5.0.0 version of Verovio!

Quote With dynamic music editions, we no longer need to translate vocal music from the 15th century intomodern notation for it to be sung by more than just a few experts. Modern technologies can help usmake early music in its original notation accessible to a broader audience.

Prof. Dr. Anna Viktoria Katrin Plaksin

Vocal Polyphony of the 15th and 16th Centuries

Mensural Notation

Mensural notation is a type of music notation used from the 13th to the 17th century, particularly for polyphonic vocal music. While medieval neumes did not indicate note durations, the spread of polyphonic music made it necessary to precisely represent the rhythm of a piece. Mensural notation was developed to capture the complex rhythmic structures of polyphonic music, but it does so in a different way than modern notation. In modern notation, the duration of a note is clearly defined by its shape and additional symbols like rests. In mensural notation, however, the duration of a note depends not only on its shape but also on its immediate context. This means that the value of a note can change depending on the notes before and after it.
Mensural Notation without consideration of contextual notelengths
The same music with the right consideration of contextual notelanghts and rythm simlifies working with early music.

Today, reading and performing from mensural notation requires specialized knowledge. As a result, this repertoire has traditionally been transcribed into modern notation by experts. However, translating from one music notation to another is much like translating books or films: certain details cannot be fully captured and are inevitably lost in translation. We are working to make this untranslatable contextual knowledge accessible and preserved through modern technologies.

Choir Books, Partbooks, Scores

Musical sources from the late Middle Ages and the Renaissance are mostly notated in the so-called choir book format. In this format, the different voices of a musical piece are spread across a double page, often in separate sections or on different pages. This means that each voice—such as soprano, alto, tenor, and bass—stands alone, and singers only have their own part in front of them. Particularly in the 16th century, printed music was often produced in the form of partbooks, where each voice was contained in a separate book.
Every Voice on its own

Scores, on the other hand, became the standard format in the 17th century. Unlike choir books or partbooks, a score contains all the voices of a piece simultaneously, arranged one below the other. This allows for a clear view of the musical relationships and the interplay of all voices at a glance, which is especially useful for conductors or composers. While the choir book format was primarily intended for performance, the score is better suited for gaining an overview of the entire work and understanding the harmonies and rhythms between the voices.

Towards Dynamic Music Editions

In recent years, the MEI format (Music Encoding Initiative) has gained significance due to its versatility and functionality. It allows for the detailed encoding of musical notation, enabling the digitization of music documents in a machine-readable form. It is the only format that supports various historical notations, such as neumes, lute tablatures, and mensural notation, making it particularly valuable for early music applications where historical sources can be digitized in an authentic form.

Verovio is one of the most widely used software libraries for displaying musical notations in the MEI format, playing a central role in digital music editions and various tools. For mensural notation, the Measuring Polyphony Editor, MuRET, and MeRIT are particularly noteworthy. This project aims to enhance the support for mensural notation in Verovio.

Code in MEI-format

The MEI format already supports both the organization of a musical piece in separate voices and as a score. However, Verovio currently lacks the ability to handle MEI files in separate voices and automatically transform them into a score. The context-dependent note durations in mensural notation present an additional challenge, which has been addressed by Martha Thomae’s “scoring-up” algorithm. This algorithm calculates rhythmic values based on context and will now be implemented in Verovio to enable the automatic calculation of note durations and the dynamic switching between individual voice views and score views. This will allow for the creation of rich digital music editions with multiple view modes, where users can easily switch between the original voice-based presentation, individual voice views, and an automatically generated score. The vertical arrangement in a score view also allows a basic understanding of a piece in mensural notation, even without in-depth knowledge.

Additionally, the improved support for mensural notation holds great potential for optical music recognition (OMR) systems. These systems can automatically extract musical content from historical sources, though currently only on the basis of individual voices. Integrating the new functionality into Verovio will enable the automated processing into a score and dynamic display as individual voices or a score, optimizing the entire digitization process.

Through this project, we have strengthened digital applications in the field of early music, making the repertoire more accessible to students and non-experts.

This project is funded within the 3rd Forum „(Further) development of Research Tools & Data Services in NFDI4Culture“ by NFDI4Culture (DFG project number 441958017).

Source of the examples:

Du Fay, Guillaume: Salve, Regina misericordiae, orig. “Wilhelmus duffay.”, copy, Bayrische Staatsbibliothek München, Mus.ms. 3154, f. 86v-87r