Selective Encoding: Reducing the Burden of Transcription for Digital Musicologists
Authors: Saccomano, Mark / Rosendahl, Lisa / Lewis, David / Hankinson, Andrew / Kepper, Johannes / Page, Kevin / Shibata, Elisabete
Date: Wednesday, 6 September 2023, 9:15am to 10:45am
Location: Main Campus, L 1.202 <campus:note>
Abstract
One of the largest barriers to digital musicology is the time required to create an encoded music file. These files contain the musical data that often serve as the primary material in digital research. While tools exist to automate parts of the process, (format converters for music notation software, for example, and optical music recognition technology) most of the symbolic content–pitches and rhythms–still needs to be entered manually, note by note. Even for an experienced encoder, it can take up to two weeks to finish encoding a single symphonic movement.
To facilitate the creation of corpora for digital analysis, we have developed a procedure for encoding only the portions of a score relevant to a particular study. These encodings can then be extended at a later time, and by any other scholars who have access to them. Currently, there is no standard way to record metadata that details which specific aspects or sections of a score have been encoded. This paper will introduce a pair of possible methods, constructed in the course of our research and tool development, to enhance the ability of MEI to accommodate these partial encodings and then evaluate the benefits of using one approach over the other.
The first method takes advantage of MEI’s capacity to create customized schemas. It features both new and repurposed elements and attributes to model a music encoding as the complete collection of a source’s encoded and unencoded sections, with each section having a unique identifier and a set of parameter values that allow for easy-to-find and easy-to-read metadata regarding the extent of the score’s digital representation.
The second, simpler method works within current MEI structures and consists of additional documentation to clarify existing usage, taking advantage of element entailments along with text descriptions to identify and retrieve encoding information from an MEI file.
We present our experience in creating the corpus for the Beethoven in the House project, a digital study of Beethoven’s symphonic works arranged for performance in the home. Our project serves as a case study that illustrates some of the assumptions that underlie these two methodologies, and how project-based considerations can lead to the adoption of one approach over another. Time and resources were limited, and encoding entire symphonic scores was neither feasible nor necessary, as musicologists on the project were only interested in comparing certain passages of the large scale works and their realization in various arrangements. This led to our interest in developing and documenting methods in MEI for creating what we call “selective encodings.”
We discovered that the introduction of new elements and attributes in MEI can be at odds with an archival philosophy that prioritizes the preservation of materials and interoperability of digital resources. And while a project may have needs that can be easily addressed with additional data structures, it can be worthwhile to consider instead adapting to the data model of an existing standard, thus better ensuring that a project’s research contributions can be shared and its data reused.
About the authors
Andrew Hankinson is a researcher and software developer for the RISM Digital Center in Bern. He has held positions on the technical group and board of the Music Encoding Initiative.
Johannes Kepper studied music and media science as well as computer science at the Musicology Seminar Detmold/Paderborn and the University of Paderborn. Since 2006 he has been active in the development of the Music Encoding Initiative (MEI) and is the German PI of the Beethoven in the House project.
David Lewis trained as a historical musicologist at Kings College, London. He is currently a Researcher at the University of Oxford e-Research Centre and lecturer in Computer Science at Goldsmiths.
Kevin Page is an associate faculty member and senior researcher at the University of Oxford e-Research Centre. He is co-founder of the Digital Libraries for Musicology conference, teaches digital musicology and linked data methods for the Master’s programme in Digital Scholarship at Oxford, and is the UK PI of the Beethoven in the House project.
Lisa Rosendahl is a research associate on the project Beethovens Werkstatt at Beethoven-Haus Bonn. With master’s degrees in history and musicology, as well as a certificate in digital humanities, she brings an interdisciplinary approach to her research on music and social history of the 18th and 19th centuries.
Mark Saccomano is a music theorist at Paderborn University and a post-doctoral researcher for the Beethoven in the House project. He previously taught music history and music theory at Columbia University in New York and was adjunct professor of music at Montclair University in New Jersey.
Elisabete Shibata focuses on the connection between music and new technologies. She is currently pursuing her PhD under Prof. Dr. Frank Hentschel at the University of Cologne, where she is investigating the digital representation of arrangements using Beethoven’s music as an example.
Contribution Type
Keywords