ALT-text/commentary/NABC colliding with note accents r1-r8 and pitch extremes

[lbssousa]

Gregorio version: 6.2.0-develop (built from develop at commit 7c806e0)

Summary

The extra vertical space GregorioTeX reserves above/below the staff (to avoid collisions with above-lines-text, gregorian commentary, the NABC lines, and the lyrics) is driven entirely by each glyph's highest/lowest note pitch (compute_height_extrema() in gregoriotex-write.c, propagated to Lua via \GreGlyphHeights, consumed by adjust_additional_spaces() in gregoriotex.lua). Two glyphs that are drawn beyond the note's own pitch aren't accounted for by that computation, so they can collide with neighboring elements.

r1-r8

accentus, accentus reversus, circulus, semicirculus, semicirculus reversus, musica ficta flat/natural/sharp can collide with above-lines-text/commentary/the upper NABC line

These are always rendered above the note (at least at pitch highest_pitch - 2, see \gre@vepisemaorrare in gregoriotex-signs.tex), but compute_height_extrema() never looks at note->special_sign, so this extra height never reaches the space-reservation logic.

MWE (dimensions reduced only to make the collision visible without a huge score; the underlying gap exists at any setting):

\documentclass{article}
\usepackage{gregoriotex}
\grechangedim{abovelinestextheight}{0cm}{scalable}
\grechangedim{abovelinestextraise}{0cm}{scalable}
\begin{document}
\gabcsnippet{<alt>al</alt>ta(gr1)}
\end{document}

Screenshot:

Low notes (pitch a/b) can collide with the lower NABC line

name: low-pitch-blnabc-spacing;
nabc-lines: 2;
%%
(c4)
gra(a||ta)

Screenshot (the ta neume sits immediately against the low note, while the lyrics below have a generous gap):

blnabc_lower in adjust_additional_spaces() (gregoriotex.lua) uses a fixed dimension
(belowlinesnabcheight) and never grows with additional_bottom_space, unlike the lyrics position,
which already does grow with it. So as a note gets lower, the note-to-blnabc gap shrinks while the
blnabc-to-lyrics gap grows.

[davidweichiang]

Regarding the accents, I can imagine two solutions:
A) make these accents increase the height extrema by an integer, as is done for episemas. This would be the minimal change.
B) always use the actual height extrema of the notes and signs. This would result in a simpler implementation than the current one. But it would change existing scores.

My question is about the decision to make the additional vertical space always an integer multiple of the line spacing: was this for aesthetic reasons or was it easier to implement at the time?

[davidweichiang]

Regarding the NABC, my recollection is that collisions were also possible above the lines, and I kept the existing behavior for below the lines.

[davidweichiang]

The NABC issue was discussed in #1662 and it was stated there that we would add a belowlinesnabcthreshold when the time came that someone needed it, and I guess that time is now!