Our impact

Intensity of Scientific Activity: Highlights of Research Achievements During the FRQSC Emerging Team Grant (2014–2015)

Cantus Ultimus (Content Axis)

The Cantus Ultimus project applies the latest OMR technologies to plainchant manuscripts in order to transform the existing CANTUS database (directed by Lacoste) of nearly 400,000 chant records into a state-of-the-art research environment in which both music and text are fully searchable. Within the past year, we have processed and made publicly available online two of the oldest surviving chant manuscripts, the late 10th-c. St. Gall manuscripts (CH-SGs 390 and 391), and completed the OMR of the Salzinnes antiphonal.

New Version of Diva (Content Axis)

Diva is a web-based, open-source digital document viewer, developed by Hankinson (postdoctoral researcher), Wendy Liu, and Evan Magoni, and managed by Fujinaga and Pugin. Diva was designed for websites of libraries, archives, and museums so that they could present high-resolution images of documents in a user-friendly interface optimized for speed and flexibility. The new version, released in August 2015, supports the International Image Interoperability Framework (IIIF). The IIIF is an important new initiative committed to developing a set of common interfaces that support interoperability between image repositories, facilitating horizontal integration across libraries and archives all over the world.

Improved ELVIS Database (Content Axis)

The ELVIS Database (https://elvisproject.ca/) is an open, crowd-sourced database of music in symbolic notation, maintained by MIRAI. This summer, undergraduate student Alex Parmentier improved the search capabilities of the database, making it more powerful and adaptable, and added new functions allowing users to upload new pieces and modify pieces they have uploaded. In collaboration with our project manager and graduate students in musicology, he clarified guidelines for data entry and made the interface more attractive and user friendly. Our continued building of the database (now totalling over 6,000 movements and pieces) has made large-scale corpus studies possible.

Increased Flexibility of the VIS Counterpoint Web Application (Searching & Analysis Axis)

Our open-source software for analyzing counterpoint, VIS, and its corresponding web application (https://counterpoint.elvisproject.ca/) make large-scale corpus studies accessible to music theorists and musicologists without programming experience. Within the last year, we have made significant improvements to the stability and flexibility of VIS. In summer 2015, Ryan Bannon (undergraduate student and lead programmer on VIS) began integrating VIS into Rodan, a workflow engine developed by Hankinson. This integration involved the low-level code integration of VIS tasks into Rodan such that they can be realized as individual atomic workflow tasks, and the development of a user-friendly web application that allows researchers to generate VIS-based workflows via a graphical user interface. We are on schedule to release our web application in spring 2016.

New Version of jSymbolic (Searching & Analysis Axis)

jSymbolic is a software tool developed by McKay for analyzing symbolic music files by extracting a range of characteristic statistical information (called “features”) relating to musical elements such as pitch, rhythm, harmony, instrumentation, dynamics, and texture. In 2015, McKay and Tristano Tenaglia (undergraduate student) developed a new version that can extract features from MEI files (Music Encoding Initiative: an open-source, computer-readable music encoding format) and from windowed sections of a score (rather than the score in its entirety). Additionally, jSymbolic is now able to export features into WEKA ARFF, which will be useful for machine learning, facilitating such tasks as composer, style, or genre recognition.

jMei2Midi (Searching & Analysis Axis)

McKay has also developed a new Java application and library called jMei2Midi, which can convert MEI files to the widely recognized MIDI file format. This allows us to process music accessible only in MEI with software not yet able to read MEI files. jMei2Midi’s general parsing libraries will also provide a useful resource for developers in the process of incorporating MEI-parsing capabilities into their own software.