Interpolation issue – glyph "A"

[lucaa (luca arnaud)]

Hi there!

In the making of a multiple master font (weight & width), there is an interpolation issue that I can’t understand. I drew the uppercase A with a break on the vertical strokes (in order to enlarge the inside counter on bolder weights). As I don’t need this enlargement with lighter weights, I simply overlapped the two points (to keep master compatibility). But the instances computed with the light weight while widening the A (preview at the bottom part of the screenshot), are emphasizing the break.

It does not looks like a bug ; could it be the great change of angle of the strokes due to the extreme widening ? I still need this break in order to enlarge the bolder weights. I already checked master compatibility (number of points, order of points, order of paths).

Has anyone experienced a similar thing ? those are pretty common letterforms

[SCarewe (Sebastian Carewe)]

Do the instances shown in the bottom panel correspond in weight to your masters? It looks like they might be lighter and thus represent an extrapolation.

[SCarewe (Sebastian Carewe)]

Sorry, even if that were the case, this is a separate issue. See here:

Variable straight slanted shape bug

For the relative angle of two lines not to change in interpolation, either the angle must stay the same, or the ratio of both line lengths must stay the same. It is better known when curves produce a kink in interpolation, as shown in “Designing Multiple Master Typefaces” (Adobe 1995, 1997), but is also true for lines. [Bildschirmfoto 2022-10-15 um 17.54.48] The problem is that not only the slant of the letters changes, but also the top bar of the T gets fatter.

Do you, by any chance, have “font type: Variable” set in Font Info > Other? Set it to default.

[lucaa (luca arnaud)]

Thanks a lot! The message you’ve quoted references the Adobe guide for MM fonts, and it explains the issue (see below, b and c).

I’ve arranged the break’s points relative height (to make it even from ultra condensed to ultra wide) and it solved the issue.

(and no, it wasn’t extrapolation : the extreme instances corresponded in weight and width to my masters)

Thanks again

edit: I tagged the topic as “resolved”, is it the right way to proceed ?

[FlorianPircher]

@lucaa

edit: I tagged the topic as “resolved”, is it the right way to proceed ?

You can add the tag, but you don’t have to. We tend to keep topics open-ended in case something new comes up.