Adair Chantbook Project

Phase 2: 2020-2024

Abstract

This phase of this project had three goals: 1) to use a home-built, multispectral imaging system to perform detailed study of palimpsests in a medieval chant book owned by LaBudde Special Collections at the University of Missouri Kansas City Libraries; 2) to develop an alternative, deep learning model that will allow us to derive visible light multispectral images from normal RGB images; and 3) to apply both direct and indirect methodologies comparatively, testing their efficacy on the UMKC chant book and on two additional heritage documents held by other libraries in the Kansas City region. We speculated that if this alternative be viable, it would dramatically reduce the cost and lower the economic barriers for other scholars, archivists, and librarians who would like to use multispectral analysis for their materials. In effect, it would allow humanities centers, libraries, and archives the ability to conduct their own investigations with these techniques using readily available and affordable equipment.

The primary test case for this project was a handwritten codex known as the Adair Chant Book held by the University of Missouri-Kansas City libraries. This Spanish book contains chants inserted from the 15th to the 17th centuries CE. The monks of El Convento de San Pedro Regalado, a monastery located in northern Spain, owned the book in the early twentieth century, but its original owners are unknown. At some point, the book was refashioned, with a number of manuscript leaves inserted, as well as new chant melodies. The numerous repairs, excisions, and additions made during the period of the book’s active use made it an ideal test case for understanding how multispectral imaging can enhance book history. Entire chants, both notation and lyrics, have been scraped away and written over with new ones, creating a series of palimpsests. Many of the changes appear to be responses to liturgical requirements of the Council of Trent (1545-1563). The introduction of an unidentified choral mass associated with the family name Strozzi offers a key text by which expected to learn more, but that avenue remains inconclusive. Multispectral imaging did help uncover several of the mysteries behind the chant included, but traditional codicology also helped determine some rationale for the book’s present form.

The LaBudde Special Collections at the Miller Nichols Library has a virtual exhibit that draws on a transcription and concert performed by Dr. Janet K. Kraybill for her Doctor of Musical Arts Degree in the UMKC Conservatory in 1999-2000.

Results

• An image gallery contains RGB, MSI, and SSR images, as well as a zoom feature, for a comparison of the various technologies.
• A white paper describes the development and successful testing of the SSR deep-learning model.