op63_no02: Clef placement and incorrect rests rendering

[bel28kent]

op63_no02 has several passages of cross-staff beaming that also have incorrect clef placements and rests rendering in VHV.

The relevant passages are mm. 9, 11, 13–15, 26, 28–29. Here are those passages myanked in VHV.

Here is a description of the issues in m. 9; the same issues apply to all measures listed above.

1. Rests that are rendering with quarter notes should be eighth notes (as encoded).
2. Stems on eighth notes following the septuplet runs are not rendering.
3. In measures with a bass clef change at the end, the clef change should be before the barline, as encoded. The last three notes should still be in the treble clef. This issue may relate to the cross-staff beaming. I have to encode these measures in *staff1 because of the alla ottava, which seems to be attached to the staff of the spine it is encoded in. So I have the notes encoded in *staff1, but the tandem interpretations for clef changes are still placed in *staff2. Here is an attempt to do the opposite; the alla ottava does not render.

[bel28kent]

N.B.: My placement of the treble clef in the VHV image is correct; I am using an urtext edition that differs slightly from the score I pulled off of IMSLP.

[craigsapp]

The basic problem is that XML cannot overlap element hierarchies. In MEI, <tuplet> is a container for notes, and <beam> is a container for notes. In this problems there is a tuplet which contains a subset of notes in a beam (or a beam which contains a subset of notes in a tuplet). So there are notes that need to be in both tuplet and beam, but not all notes in the tuplet/beam are in the other respective container.

The solution for this situation in MEI is called a tupletSpan, which is implemented in verovio, but it is having a problem rendering the distilled example that I created to test:
rism-digital/verovio#3500

Also, this will be somewhat complicated to implement (at least in terms of concentration), so I will probably look at it in the Autumn (and after the above issue is fixed in verovio, or a demonstration that I formulated the example incorrectly).

The octave line has complications because the notes are cross-staff (as well as the octave line), so I will also have to deal with that sort of case as well in the Humdrum-to-MEI conversion.

Probably the clef problem will be fixed after the tupletSpan is implemented.