Vertical character spacing in Chrome variable

Viewed 43

I converted an SVG font to OTF using FontForge. I know the original font has certain errors, but really don't think any of those errors would cause the following issue:

When using CSS writing-mode:vertical-lr; text-orientation:upright, I get the following results. The Firefox rendering is perfect, but for some reason the results seem to kern certain letter combinations in Chrome (ver. 96). I've tried doing everything I can think of in FontForge as far as clearing kerning tables, toggling options for "old kerning" and "microsoft kerning", etc. Nothing seems to make any difference to the result.

I know that upright orientation is not well supported by browsers in general, but it's apparent that Arial font renders perfectly under Chrome, so I think there must be something I can do to fix this problem.

The font is generated in FontForge from an SVG font that uses vert-adv-y. The values for that parameter are correct in the SVG (the SVG has other errors, but I don't think they have any bearing here). I'm linking a copy of the exported OTF file. Perhaps someone may discover some sort of kerning or letter-advancing issue there.

Or perhaps Chrome is doing some sort of automatic kerning, in which case, why doesn't it happen to the Arial font? Maybe the Arial font has GPOS settings, whereas my font has old-style kerning? I really have no clue.

LINK TO DOWNLOAD OTF FILE. Although the letters on the left of the image are lowercase, they are produced using the uppercase keys A, B, C, etc.

my vertical font on left; Arial on right

1 Answers

Okay, so this seems to have been one of those miraculous-seeming instances when a problem seemed to fix itself. I resaved the file, which then kicks it to a microservice running FontForge, which coverts the SVG to an OTF, and when it reloaded everything appeared good. I still have no idea why, since I didn't update the microservice, and I haven't changed the SVG-generating code AFAIK. All I did was upgrade Chrome (and the new version had been showing the issue with the old file), but it should still be sending the same data. Still, I'm not complaining!

Related