Abstract
A common task in the humanities is to explore the genesis of textual or musical works. While the creative processes behind a composition are usually not directly accessible, they can sometimes be inferred from traces left on scrap paper. For genetic editions, such traces are crucial for understanding the material related to a composition and identifying intermediate steps in the creation of the work.
Sketchbooks are challenging for several reasons. As personal documents not intended for an audience, they are often illegible and incomplete, omitting many musical parameters found in standard notation. For instance, a series of noteheads with stems might simply indicate a sequence of pitches rather than specific rhythms. Moreover, musical sketches often present material in forms that differ greatly from its eventual implementation—for example, an orchestral passage might be reduced to one or two staves—making it difficult to relate sketches to corresponding passages in the final work. This is further complicated by the fact that sketchbooks may contain ideas for multiple works composed simultaneously, not all of which are realized, introducing significant uncertainty and requiring interpretation on many levels.
Modeling such material and its editorial treatment for digital editions involves considerable complexity. Following the TEI’s model for genetic editions, an encoding model for music documents employs multiple layers of representation. The most basic level is the identification of the visual appearance of shapes on the page. This is followed by their interpretation as musical symbols and contextual editorial supplements to establish musically comprehensible text. Finally, these are linked with complete works where feasible. While this ensures transparency of editorial processes, creating and maintaining such interlinked data is demanding.
Our paper introduces a prototype software, FacsimileExplorer, which assists in creating encodings for digital genetic editions of musical sketchbooks. It implements a data model that combines SVG and multiple variants of MEI.