LaTeX \bar not centered over 'v' and 'w'

[legendre6891]

In the STIX2Math font, the math snippet $\bar v$ and $\bar w$ gives the following result, in which the vertical bar is not correctly centered.

[jagd]

\hat \tilde etc. have this issue, too

[jagd]

As stated in the title, it seems the italic v and w may be the most affected ones. Besides, the kerning of italic j could also be optimized.

[juw]

For \tilde{b} as for example used to write the supersymmetric partner of the bottom quark the tilde also seems to be placed to far to the right: MWE:

\documentclass{standalone}
\usepackage{mathtools}
\usepackage[math-style=ISO,bold-style=ISO]{unicode-math}
\setmainfont{STIX Two Text}[Numbers=OldStyle]
\setmathfont{STIX Two Math}[RawFeature=+cv03;+ss08]
\begin{document}
$\tilde{b}$
\end{document}

[jagd]

My temporary solution is to remove the kerning of math accents, something like

\let\orghat\hat
\def\hat#1{\orghat{\kern0pt #1}}

also to \bar, \tilde etc. Then v, w, m, g and their bold variants are saved.

The math accents on the italic A and f would be non-optimum but acceptable.

[jagd]

Related issues (see unimath-symbols.pdf, the green color is STIX2)

[tiroj]

This has been left to last in the v2.20 MATH table revisions. I will try to address the most problematic / most obviously incorrect things this week, but in some cases there may be preferences regarding positioning to be discussed and addressed via future revisions, e.g. the comment above re. the tilde on the italic b: it is optically centered on the slant of the base character width, but perhaps users would prefer it over the ascender of the letter?

[tiroj]

I have been stepping through the Top Accent Attachment data in the MATH table, and I am not seeing anything that reflects the results shown in the screenshots in this issue. This is the positioning shown in the MS MATH table editor:

[tiroj]

Have been through entire Top Accent Attachment data in MATH table, making a few small adjustments, a couple of corrections, and adding the newest bold italic variants.

Closing this issue for now, but will re-open for re-testing when v2.20 is available. Interested to see comparison testing with other layout engines and platforms.

[tiroj]

Updated sources and fonts delivered to STI Pub for testing. Issue reopened pending public release.

[Firestar-Reimu]

I found this issue too on v2.13

[tiroj]

?

The positioning of the marks as shown in that image are correct.

[Firestar-Reimu]

Sorry, but I prefer this style (from XITS, and TeX Gyre Termes Math is just like this)

Or this from Lucida Bright Math OT

---

For example, letters which have ascenders like b, d, h, k, should have the accents above the ascender (not the middle of the letter).

[tiroj]

This is in part a matter of taste, and there’s no consensus on the ‘correct’ way to position marks relative to ascenders and descender strokes. I’ve made fonts with one or other approach, informed by the intended use of the fonts, preference of particular user communities or clients, as well as by practical considerations. There is, apparently, some empirical evidence that optically centering the marks on the letter width is better for legibility. I’ll confer with @davidmjones and see whether any of the STI Pub members have an opinion on this.

Note that this question is distinct from the original issue.

[Firestar-Reimu]

For \tilde{b} as for example used to write the supersymmetric partner of the bottom quark the tilde also seems to be placed to far to the right: MWE:

Yes I have the same problem on \tilde{h}

the tilde is better to be on top of the ascenders.