MEI Encoding

The data produced by the previous steps is still not in the MEI format. The role of the MEI encoding job is to transform this data from JSON into MEI. This obviously requires the data itself, but also requires a CSV file mapping the class of a glyph in that manuscript (e.g., neume.punctum) to an MEI fragment that represents it (e.g., <neume><nc/></neume>). The easiest way to produce this kind of CSV file is through the MEI mapping tool, although it can also be manually created through a tool like Microsoft Excel or LibreOffice Calc with careful formatting.

If text alignment data is provided, the MEI Encoding job will also attempt to match the text to neumes and group them as syllables. Otherwise there will be one syllable in the entire staff. For grouping neume components into neumes, a heuristic is used based on the position of the glyphs.

Correction with Neon

Neon renders the MEI produced in the previous step over the source image by using Verovio. A full description of its features are available on the project’s wiki, but there are a few key points for using it as part of the workflow.

  1. Adjust staff size and rotation. This will help glyphs appear over the correct place on the page.
  2. Correct elements that were incorrectly classified. A typical case is that a C clef is classified as two neume components, or an F clef as three.
  3. Correct the pitch of elements. This is done by just dragging to the correct position.
  4. Correct the grouping of neumes and syllables. If text alignment data was provided, this is often close to correct already.
  5. When you’re done, press the “Validate” button in the upper left of the screen to send the new MEI file back to Rodan so it can be saved as a resource.