Cross-staff articulation vertical positions

[craigsapp]

Something happened quite recently to cross-staff articulations, where 3.10dev is colliding some of the staccatos with the cross-staff beam:

Also notice that the clef-beam collision in 3.10dev which is not in 3.9. This is probably related to the stem on the last note of the cross staff note being too short (probably allow stems to be shorted to 4 VU at their shortest when in a beam). The staccatos seem to be tool close to the beams in 3.10dev, and I like the staccato spacing on the cross staff beam in 3.9 the best.

The staccato on the last note in the cross staff beam is too low in both cases (should be in the space just below the note instead of two spaces below).

Click to view MEI data for the above example

```xml Verovio

Transcoded from Humdrum

K. 84 Scarlatti, Domenico Sonata in C minor, L.10, K.84     Domenico Scarlatti 1685–1757

```

[craigsapp]

The articulations like to move around based on layout, so they don't exactly know where the beam is located:

The angle is right, but most likely the beam was moved after the positions of the articulations are assigned.

It could be related to spacing (such as stretching the music after castoff). Here is where I start with a tight spacing with the same music:

In this case the staccatos on the top beam in the first measure are also being affected, even though it is not a cross-staff beam.

Here is a wide spacing where they are correct:

[brdvd]

@craigsapp Unfortunately, I cannot reproduce this. Looks ok on my side:

Can you fill in more details how to reproduce this issue? Which rendering options do you use?

The change in behaviour appeared after #2699 .

[craigsapp]

It is also working for me now, with the staccatos not changing positions, and the staff distance remaining constant at various spacings of the music layout:

There is a strangeness in that the staccato marks on the cross-staff beam are further away than the regular beams.

Cross staff staccato placement:

Compared to regular beam placement where the staccatos are closer:

[craigsapp]

After compiling the most recent PR #2905 to fix the problem, it is still there (the original problem mentioned at the top, not the staccato spacing fix).

It is subtle to produce, and requires more musical content to make it visible.

Here is a demo with the entire score:

https://verovio.humdrum.org/?file=scarlatti/sonatas/L010K084.krn#mh11

(this view may depend on the width of your browser when viewing the above link).

When I make the musical spacing slightly larger:

https://verovio.humdrum.org/?file=scarlatti/sonatas/L010K084.krn&k=w#mh11

The problem becomes visible:

It is not directly related to spacingLinear/spacingNonLinear, since I can resize the window with the same spacing and the staccatos are fixed:

It seems to be related to the justification process, and in particular when the measure has to be stretched out a lot to fill in the system (but if there are only 1 or 2 systems in the the problem does not show or is hard to make show).

Here are the first 14 measures of the test music (measure 11 being the measure of interest):

Click to view MEI data for above example (bars 1–14)

```xml Craig Stuart Sapp 2021 Craig Stuart Sapp Licence: (CC BY-NC-SA 4.0) https://creativecommons.org/licenses/by-nc-sa/4.0 Verovio

Transcoded from Humdrum

Encoded by: Kelly Leistikow

Version: 2020/12/21

K. 84 Scarlatti, Domenico Sonata in C minor, L.10, K.84 Domenico Scarlatti: Opere complete per clavicembalo (Milan: Ricordi, 1906-1913)     Craig Stuart Sapp, ed.   Domenico Scarlatti 1685–1757

Allegro ( = 152) f p p p p f p p sf p

```