Andrea Puentes-Blanco presents a detailed account of the proceedings from The Digital Musicology Workshop, held at the Oxford e-Research Centre from 20-24 July 2015. Andrea is a Ph.D. candidate at the University of Barcelona and the Spanish National Research Council (CSIC) working on sixteenth- and seventeenth-century manuscripts of hispanic polyphony and interested in applying digital technologies to her research.
The Digital Musicology Workshop was held at the Oxford e-Research Centre (University of Oxford) from 20 to 24 July 2015, organized by Kevin Page (Oxford e-Research Centre) as part of the Digital Humanities at Oxford Summer School. This five-day workshop consisted of a series of lectures and hands-on sessions to offer an introduction to computational and informatics methods that can be, and have been, successfully applied to musicology. Participants from around the world were mainly Ph.D. students and scholars in musicology, but also in computer science and music librarianship. The workshop was introduced by Tim Crawford (Goldsmiths, University of London) and J. Stephen Downie (University of Illinois at Urbana-Champaign). Crawford described how he became involved in the field of digital musicology: he described his background as a lutenist and how the principal features of lute repertoire (mostly anonymous and borrowed or quoted from other music) led him to apply computational methods to study this repertoire in order to discover more of the music and establish concordances. Talking about Music Information Retrieval (MIR), Downie presented the work of ISMIR (the International Society for Music Information Retrieval) and how this interdisciplinary research community works to create new algorithms to improve the process of retrieving information from music.
One of the major topics of the workshop, ‘big data’, was introduced by Stephen Rose (Royal Holloway, University of London). With reference to Franco Moretti’s ‘distant reading’ (2005) for literature studies (understanding literature not by studying particular texts, but by analysing large literary corpora), he reflected on how working with big data can open a new and larger perspective for music history. Rose described the Big Data History of Music project carried out at Royal Holloway in collaboration with the British Library. Working with large amounts of data from some of the world’s biggest collections of published music, music manuscripts and concert programmes (including RISM and the British Library’s catalogues, among others), this project aims to develop new methods for research in music history through statistical analysis and visualization of this data. Rose talked about the importance of preparing and cleaning data in any project that works with large amounts of data, and he showed us some examples and suggestive findings of the project concerning music publishing in the period 1500–1700.
Continuing with big data, Downie, David De Roure (Oxford e-Research Centre) and Ichiro Fujinaga (McGill University, Montreal) presented the ambitious SALAMI (Structural Analysis of Large Amounts of Music Information) project, hosted by IMIRSEL (International Music Information Retrieval Systems Evaluation Laboratory) in collaboration with the University of Illinois at Urbana-Champaign, McGill University, the University of Oxford and Queen Mary University of London. This project analysed the general structure of large sets of music recordings from the Internet Archive using a range of computer algorithms designed to detect musical structures; the accuracy of these algorithms were tested against human annotators (graduate students). The project’s aim was to produce a large web-accessible corpus of analyses of several hundred thousand recordings including classical music, jazz, folk and world music. This innovative project has opened up new perspectives on music analysis, a discipline traditionally conducted by individuals on a small scale.
Kevin Page introduced another important topic: the Semantic Web and Linked Data (a new way of using the Web to publish highly interlinked and also machine-readable data). He described the principles of these technologies and commented on the potential of using Linked Data to publish, reference and reuse the output of digital music research. Carolin Rindfleisch (University of Oxford) presented excellent work in progress, related to the Semantic Web, from her Ph.D. dissertation. Her objective is to undertake a systematic analysis of the reception and interpretation of Wagner’s leitmotifs in Der Ring des Nibelungen at different times and in various cultural contexts. She is developing a Semantic Web ‘ontology’ (a description of concepts and their relationships expressed using RDF, the Resource Description Framework, which is one of the foundational technologies of the Semantic Web) which allows her to express the complex relationships and influences between different interpretations of the leitmotifs in a structured way.
Richard Lewis, David Lewis (both Goldsmiths) and David M. Weigl (Oxford e-Research Centre) ran two very useful hands-on sessions about digitized notated music formats (MEI and MusicXML) and Music21, a toolkit developed by Michael Scott Cuthbert at MIT (Massachusetts Institute of Technology) which can be used to search for patterns and to prepare reproducible music analytic tools. In order to work with Music21 we also learnt some basic aspects of Python, the programming language used by this tool.
Two more hands-on sessions were about audio analysis and machine learning to search for patterns in large collections of audio. Christophe Rhodes (Goldsmiths), Chris Cannam (Queen Mary) and Weigl showed us some tools for extracting and visualizing audio features (such as amplitude, timbre, tempo, pitch and so on) using Vamp plugins and the software Sonic Visualiser. Ben Fields (Goldsmiths) and Tillman Weyde (City University London) presented the Digital Music Lab project (a collaboration between City University, Queen Mary, University College London and the British Library) and the interface developed by it (Digital Music Lab VIS). Through this tool, we were able to explore, analyse and compare large audio music collections from the British Library, CHARM (Centre for the History and Analysis of Recorded Music) and I Like Music.
Fujinaga gave an overview of the history and state of the art of optical music recognition (OMR). He emphasized the importance of image preprocessing such as staff-line removal to improve optical recognition. He also described various applications in the evolution of this technique, such as GEMM (Gamut for Early Music on Microfilms), which allows OMR of music from microfilms. Joining the principles of Linked Data and the process of OMR, Crawford, Page, Weigl and David Lewis presented a case study in early music. The objective of their Semantic Linking of BBC Radio (SLoBR) project is to create a resource of data in early music which is accessible, searchable, reusable and linked with some other extant early music databases (Early Music Online and Electronic Corpus of Lute Music, both of which have been the subject of OMR research) as well as other external resources and databases including DBpedia, MusicBrainz, and BBC broadcast data.
A lecture by Julia Craig-McFeely (University of Oxford), entitled ‘Blind Alleys, Science Fiction, Redundancy and Modernization: How Musicology Is and Isn’t Evolving in Response to the Digital World’, addressed some important issues about the challenges that the world of digital musicology, continuously growing and developing, raises for musicologists. She presented a cautionary point of view regarding the world of digital musicology. She commented on its possibilities and applications: for instance, the large resources available online (musical sources, archival material and bibliography allowing for interdisciplinary research) or digital restoration techniques applied to early music manuscripts and prints (being carried out by projects such as DIAMM and Tudor Partbooks). But she also pointed out the negative side of the digital world: for example a lifetime’s work being made redundant by a digitization project.
Both lectures and hands-on sessions covered a large range of topics in digital musicology. However, I would have liked to have heard some thoughts and been given some practical information about the process of producing critical digital editions of musical works, manuscripts or corpora, since producing critical editions is, traditionally, one of the main activities of musicology, and the digital environment can offer more possibilities than the traditional hard-copy format. I hope something like this may be included in future iterations of the workshop.
Thanks to good organization, practical sessions were easy to follow: each laptop had the appropriate software installed and preconfigured and we were guided step by step in all tasks. This workshop opened to me a new and wide perspective on the potential of using computational methods for musicological research. But, above all, I have realized that the world of digital humanities is a collaborative one. Humanities research should no longer be a lonely activity; we must find ways of working with people with a technical background, because scholars working alone in the non-digital world cannot, in most cases, achieve as much as those working in the digital sphere.
During the entire week and in the final round table, delegates were encouraged to make comments and ask questions about the projects and technologies being presented and also about their own research. In sum, it was an excellent chance to expand our digital knowledge applied to music research and to stimulate our ‘digital imagination’.
If you or someone you know would like to write a post for the RMA Student Blog then we’d be happy to hear from you! Please get in touch with one of your student representatives.