How can digital methods help reduce the complexity that arises from differences among multiple sources of a musical work? We present a web-based application designed to visualise differences between digital representations of various versions of musical texts. This approach addresses challenges similar to those encountered in fields like comparative genomics.
Our tool offers two visualisation methods:
-
a dynamic network visualisation (Linhares et al. 2017), where nodes represent sources and edges represent differences derived from pairwise diff files, and
-
a Neighbor Joining tree (Saitou & Nei 1987), which reconstructs a tree structure from a distance matrix. By aligning diff-files (Hunt & MacIlroy 1976) sequentially, our tool visualises how and when changes occur over time.
Both visualisations display the relationship between sources at specific moments, allowing users to trace divergence and convergence, potentially discovering hints of contamination, and demonstrate how differences listed in musicdiff files (Foscarin et al. 2019) can be explored intuitively. The (animated) dynamic network approach is especially well-suited to illustrate temporal developments, while the Neighbor Joining tree provides a clearer interpretation of distances. Each method thus offers a different analytical perspective.
By integrating both approaches into a single web application, we are making them accessible for scholarly research and digital editions alike. Beyond supporting established editorial tasks, we see particular potential in evaluating large corpora of automatically generated transcriptions using Optical Music Recognition (OMR) software. As digital editions increasingly rely on such automatic processes, tools like ours are valuable for quickly assessing how much a new transcription diverges from existing versions.
You can create the diff-files from multiple MEI-Files with the following colab: Google Colab Multiple MusicDiffs Then you can find our app here.
Presented at Third International Conference on Computational and Cognitive Musicology in Aalborg University, Denmark, 8-10 October 2025. Poster:
If you want to deploy it yourself, you can clone the repository, then you should install dependencies with
npm install
and then can start the app with
npm run dev
You can also find this project on OSF.