Early last month on the 11th and 12th of November, 2020 a subset of the TROMPA team participated in MELDfest-- a small, virtual event focusing on MELD, one of the core technologies used in the development of several of the TROMPA use cases. MELD, an acronym for the Music Encoding and Linked Data framework, is a conceptual architecture used by software developers working with music for combining the many types of media associated with music study. The idea is to be able to connect text, audio, video, scores or images that all relate to the same overarching concept through a web of meaningful, digital connections.
The event attracted attendees from University of Oxford’s e-Research Centre, Universität für Musik und darstellende Kunst Wien, Universität Paderborn, the Music Technology Group at Universitat Pompeu Fabra, Goldsmiths, Johannes Kepler University, as well as RISM in Switzerland.
Spanning two days of presentations, discussions, and planning new ways of utilizing the MELD framework in future research, MELDfest provided a touch point for many software developers facing similar problems. The day’s events were recorded and access to these recordings are available on request, please get in touch with TROMPA team member David Weigl if you are interested.
The first presentation of MELDfest was given by Kevin Page and David Lewis from the Oxford e-Research Centre. Page and Lewis are two of the three original researchers to originally have proposed the MELD framework. In their presentation, Lewis and Page discussed work they had done on the performed works of Frederick Delius in collaboration with Daniel Grimley and Joanna Bullivant from the Music Faculty at Oxford, the Villiers String Quartet, and the British Library. In their presented case study, they demonstrated how the MELD framework can be used to support musicologists in their discussion of musical performance by affording a multi-media article which includes text, images, digitized musical scores (using the Music Encoding Initiative), and both audio and video recordings.
Following Page and Lewis was a presentation from TROMPA’s Music Scholars use case. Federico Zubani presented the latest working prototype of a user interface that will allow those interested in annotating a score the ability to not only make selective annotations on elements of interest such as the notes and measures, but also features in development for annotating what others have already noted.
Contrasting to the relatively developmental stage of the ongoing TROMPA projects, the third presentation of the day showcased what can be accomplished nearer the end of a software development cycle. Resulting from the AHRC funded Transforming Musicology and its follow-on Unlocking Musicology project, David Lewis presented a musicological workflow developed with Wagner scholar Laurence Dreyfus that allows a musicologist to provide a visually pleasing interactive dashboard, supporting the easier comprehension of more complex scholarly arguments. The workspace allows users to navigate a reduction of a musical notation that does not require the musical literacy to read 19th century orchestral score, but allows the viewer to pair the musical score and the libretto together in tandem with scholarly writing. For a more in-depth description of this technology, please see this video.
Breaking from the MELD framework for use in score analysis Graham Klyne presented work on the SOFA project, which demonstrates how Linked Data can facilitate the generation of music. SOFA allows users to create new musical compositions by pre-specifying a set of musical snippets that will be played to the listener according to a set of predetermined rules. You can read more about SOFA here.
Another TROMPA project, presented by David M. Weigl, the last author and lead developer of the MELD framework shared the latest developments on CLARA, a Companion for Long-term Analyses of Rehearsal Attempts for instrumental players. Integrating live data capture, canonical recordings, and the displaying of a Verovio rendered score, the CLARA project aims to allow pianists to be able to play a passage of music in real time, then compare their performance with the note-level performance choices of pianists they may be trying to emulate. Using the MELD framework allows for the musicians to have access to multiple forms of musical information within a single suite of software.
The presentations finished with Graham Klyne of the Oxford e-Research Center sharing recent and ongoing developments on a command line interface providing a test suite for MELD applications, an initiative funded by the UK Software Sustainability Institute (SSI).
Day one came to a close with an extended discussion on the future of the MELD framework with technical plans put in place to iron out current limitations of MELD’s traversal engine, issues with rendering scores in real time, and integration with the Solid framework supporting user contributions through read/write Linked Data.
Instead of discussing past projects incorporating the MELD framework or discussing current issues, day two of MELDfest focused on the future. On day two, discussions focused on what aspects of the framework needed to be attended to within what was shared across all current and future projects, including the use of an internal graph model for the organization of data, application setup within RDF, and issues with rendering multiple views with the music digitization format used.
While the use of MELD within the TROMPA project will come to an end in the next year, the developments with MELD on TROMPA will be helping future projects. Near the end of the second day, we had two final presentations sketching future projects that will be publicized in the near future!
Both days were filled with lots of great discussion surrounding music and linked data. Please watch this space for new developments surrounding the MELD framework as it applies to the TROMPA project.