In its first two years of activity, LAUDARE project has largely adhered to the plan and expectations outlined in the application. The first four post-doctoral students were recruited. From April 13 to 15, 2023, the University of Trento hosted the first workshop devoted to “Research and Image Acquisition”, titled “Singing as a Function: A Multidisciplinary Approach”. During the workshop, research guidelines and technological strategies for data acquisition were discussed, with input from other music computing specialists. This occasion also fostered productive interactions with teams from the Universities of Innsbruck, Montreal-McGill, Alicante, and the Biblioteca Hertziana-Max Planck Institute of Rome, aimed at sharing experiences in developing neural networks tailored for the acquisition of manuscripts combining text and music. The search for new lauda sources began in October 2022. Missions have so far been conducted in over twenty Italian cities, as well as in selected libraries in the United States, Spain, France, and England. A specific case, already subject to particular expectations, concerns the regions of Southern Italy (except for Sicily), so far almost entirely devoid of attestations up to the late Renaissance. To date, three missions have been conducted in Naples and one in some cities in the province of Bari, but no results have yet been achieved, despite the strong presence of lay confraternities, many of which remain active.
The musicologist and computer scientist teams collaborated to identify a proper relational structure for a database (WP 2). A backend has been constructed on the GSSI servers, and it is now ready for data entry. The database will ensure efficient and effective retrieval of lauda sources based on metadata and on their mutual relationships.The import process (WP 3) started with the two major monophonic laudari that have survived in their entirety – Cortona 91 and Banco Rari 18 of Florence – following an analysis of the layouts with their differing divisions of 'text' and 'music' fields, carried out for the moment using a mixed approach: READ Coop's Transkribus software for the text and McGill's Neon for melodies. Successively, the MEI and PageXML files produced by this software are transformed in standard and more suitable files, including MEI and TEI, which will serve as basis for the content analysis that we are going to start in the coming phase of the project.
Representation for Neumatic Notation: To ensure interoperability, we use established file formats like MEI, TEI, and PageXML while addressing challenges specific to their use. In computer vision, annotation formats such as PASCAL-VOC (XML-based) and COCO (JSON-based) are common. JSON's advantages, including smaller file size and better readability, make it a preferred choice. We developed a customized COCO JSON format to represent the complexity of lauda sources, integrating text and music. This innovation supports experimentation with advanced machine learning models and enhances the usability of the datasets for scientific benchmarking. Annotating Lauda Corpora is a specialized, time-intensive task, and no existing software fully captures their visual and musical complexity. To address this, we created a workflow, combining tools for text and music annotation with synchronized markers.
Cross-Disciplinary Research: To address the challenges inherent to the lauda genre, our project fosters continuous, critical dialogue among musicologists, ethnomusicologists, philologists, digital humanists, and computer scientists. Epistemological approaches and practical strategies proposed by team members are collaboratively examined, refined, and implemented. This ensures that methods and insights developed in one discipline are rigorously tested and effectively applied across others. This approach integrates practical tasks and source analysis. Technical and humanistic perspectives also converge in the project. Paleographical and codicological studies of ancient lauda manuscripts and printed books reveal how these sources were meticulously designed to facilitate the transmission and use of laude. Consequently, we aim to integrate the contextual and intertextual richness of historical lauda sources into the LAUDARE relational database. This approach seeks to preserve the historical sources' modes of presenting connections and relationships, offering modern users an experience akin to that of lauda singers.
The cross-disciplinary approach extends even beyond the LAUDARE team. In May 2023, we established a permanent scientific panel to explore the relationship between music and poetry. Sharing this knowledge beyond the academic sphere will not only enhance future surveys in libraries and archives but also raise awareness among non-specialists about a genre of Italian poetry and music that was remarkably popular from the Middle Ages to the Modern Era.
