DCMLab / reductive_analysis_app

Web app for reductive analyses of scores
https://dcmlab.github.io/reductive_analysis_app/
Other
15 stars 4 forks source link

DOI

Reductive score annotation app

This is an app to facilitate music analysis using a reductive (e.g. Schenkerian) paradigm.

The app is continually under construction but may be tested here

Table of contents

The basics

The basics are as follows:

A short example

Let us load a familiar piece into the annotation app: Bach's Prelude in C Major from The Well-tempered Clavier (BWV 846). This is done by selecting the file from your computer through the file selector accessed from the "Select a score" button. Once chosen, the app should render it directly as in

This done, we can begin selecting notes by clicking them. We can also select notes as primary by shift-clicking them.

After selecting a number of notes, we can associate them using a relation, either of some specific type, or untyped. There are a number of predefined relation types with dedicated buttons and keybindings, and a textfield for entering a custom type in the expanded menu.

Let us select the first E in the upper voice as primary, and the rest of the E's in that measure as secondary:

We can do this either by clicking and shift-clicking the individual notes, and then either clicking the button marked "Repeat" or using the "+" keyboard shortcut. However, specifically for the situation that all the notes of the same pitch in a single bar should be selected and related as repeats, it is also possible to just select the primary note of the repeat and then hitting "+".

After having created a number of relations, we can choose to do a "Reduce" step (by clicking the button in the layer menu on the left) which hides the "lowest" level of edges, as described above. For example, if we have made similar relations for the other parts in the arpeggio, we can reduce the first bar of the Prelude from this:

To this:

However, this doesn't look very nice, as we only hide the notes, and not the beams and other things related to notes. We can hide stems and beams and such things with the button "Stems etc." in the "More options" menu, or the keyboard shortcut "s".

But this still leaves us with an embarrassing amount of unused space, as well as unused ledger lines. We can, however, create a new layer, which will be added in the app below the first layer, where such unused spaced can be significantly reduced, and features like note editing becomes available.

Slightly better!

Having done a number of reductions and new layers, we can end up with something similar to:

Which starts to be useful for further analysis.

By clicking the "Save" button, you will get the option to save the original MEI, any selection of additional layers, plus the graph as currently envisioned. For example, saving the above state and then loading the resulting file will show this view:

Hierarchical analysis

More exciting possibilities are open by having overlapping and interacting relations, though this is beyond the scope of this simple example. For presenting an earlier version of this work as a Late-Breaking Demo at ISMIR2020, the following illustrative gifs were produced, showcasing some more advanced ideas:

Update: In the GIFs below, reference is made to needing to delete edges in order to see relations "behind" them. This is no longer required: instead, by scrolling with the scroll wheel (up or down), the currently highlighted edge is sent to the "bottom" of the stack, showing the edges that were "hidden".

Showcase: Schenkerian analysis

Showcase: GTTM Tree

Showcase: MOP annotation

Funding and publications

The present work has been presented at ISMIR 2020 as a Late-Breaking Demo, with this extended abstract and poster.

This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program under grant agreement No 760081 – PMSB. We thank Claude Latour for supporting this research through the Latour Chair in Digital Musicology. Additionally, the members of the Digital and Cognitive Musicology Lab (DCML) have contributed valuable insights through discussions and user testing.

Building the app

The app is in /src. Assets called by the app needs to be compiled using some front-end tooling, detailed below.

Development

  1. Duplicate .env.example to .env and edit it.
  2. Run npm install (Node > 12.13) to install all the required packages and tools.
  3. Run npm run dev and open the URL returned by the CLI.
  4. After a code merge in the main branch, update the modified build (npm run build) and push it, too.

Production

  1. Duplicate .env.example to .env and edit it.
  2. Run npm install (Node > 12.13) to install all the required packages and tools.
  3. Run npm run build to compile the app. The compiled app goes in /public.

Various