Pre-Kern On?

Post Reply
Posts: 3
Joined: 23 Apr 2021

Pre-Kern On?

Post by federicoparra »

I was wondering if there could be (or you have thought of) a Pre-Kern On script or plugin based on the way Kern On analyses pairs. I explain: the general purpose of kerning should be to fix what we cannot achieve with spacing, so I guess the worst the spacing is the more pairs Kern On will generate. But what if a Pre-Kern On gives us the same warnings as Kern On but on a glyph basis? E.g. "Letter H seems too tightly spaced compared to letter a". The idea would be to have an assistant for the spacing prior to kerning.
Posts: 1
Joined: 02 Jun 2021

Re: Pre-Kern On?

Post by Avenir »

I can second that, had a similar thought. "Pre-Kern On" could be a nice spacing guidance...
User avatar
Tim Ahrens
Site Admin
Posts: 231
Joined: 11 Jul 2019

Re: Pre-Kern On?

Post by Tim Ahrens »

Good point, I have thought about spacing assistance (or even auto-spacing) for a long time but it’s surprisingly difficult to put into an algorithm. Of course, a glyph side that only has negative kerning in all possible combinations must be spaced too loosely (or too tightly if it has only negative kerning). But that’s a rare case – even if the T is spaced too loosely it will have some negative pairs such as LT so we will not detect it. Similarly, a glyph side that is kerned tighter than another glyph side in all possible combinations (e.g. comparing all pairs before e with all pairs before o) must be too loose in comparison (or vice versa). But again, that’s rare.

You can use Kern On as a tool to check your spacing if you have a “neutral glyph”, i.e. a glyph (or glyph-side) that shouldn’t have any kerning at all, by definition. In many sans serifs, there shouldn’t be any kerning before and after H, or before h or after d.

The “Pairs” drop-down on the KO panel shows you a list of pairs before or after the active glyph. By looking at the list of pairs before or after the neutral glyphs you can easily spot spacing errors (or missing model pairs, or missing special spacing). To me, that technique feels so convenient that I can hardly think of an automatism to improve it. Also, it gives me control and makes think and look carefully.

I might add an option to let the user say “This is a neutral glyph side, it shouldn’t have any kerning”.

However, if you think about seriffed designs, are there any neutral glyphs? Does the H have any kerning? How about OH or on? Do they have kerning because they “fit into each other”, or not because they are “standard pairs”? It depends on your spacing concept. Maybe for seriffed designs, one could use | (the bar) or a glyph made specifically for that purpose (even auto-generated) as a glyph that is “neutral by definition”. And how about seriffed italics? The lowercase would usually be spaced so that nn does not have any kerning even though the shapes “fit into each other”. Would that hypothetical “neutral glyph” be somehow s-shaped? Oh, but not for the uppercase, I guess. How can we possibly define which spacing is normal?

Happy to hear your thoughts on this!
User avatar
Posts: 7
Joined: 07 Jul 2021

Re: Pre-Kern On?

Post by colinmford »

I would be interested in the idea of a "Neutral glyph side, don't kern to this". Even if it's not useful in every application, it would be useful in some (like a sans H for instance), and surely it's up to the user to decide when to employ it.
Post Reply