Merging Taktpositionen with MEI with content

[anneferger]

Use case from Beethoven in the house

[kepper]

Hi @anneferger, thanks for adding this to the backlog :) The idea is to merge an MEI file with measure positions for the whole piece with another MEI file that holds the music, but not necessarily for the whole piece. There is a number of XSLTs out there that will to the merging, but there is no generic solution for this use case yet. The main challenge is that it requires some basic mapping between the two files. On the XSLT level, I think this could be done with a measure ID from both files where the alignmet is supposed to start, but this tool would greatly benefit from a simple UI that will generate these input parameters and allow a control of the expected results. I think it would be easiest to have a small meeting about this and discuss how much effort can and should go into this…

[musicog]

If I understand correctly, this use-case is catered for by mei-friend - it is possible to ingest a facsimile-aligned sparse encoding (measures-only) as produced by the Deep Optical Measure Detector and merge it with a non-facsimile-aligned note-level encoding, using measure@n identity as a matching basis, as long as certain assumptions (same number of measures, etc) hold. It's also possible to make fine adjustments (e.g., resizing zone elements) in the graphical mei-friend interface once the merge is complete.

Please have a look at the mei-friend documentation on facsimile support, particularly the section on generating zones using a skeleton file which describes this workflow -- note that at time of writing the mei-friend documentation is still very much under development and not particularly ready yet for public consumption, but that chapter happens to be more-or-less finished!