Overshoot and Slab Serifs

Choz Cunningham

I am curious how others handle overshoots with slab serifs. On the on hand I think that perhaps they should be stronger (maybe 5%) because the neighboring slabs create such much weight on the base and cap lines. On the other hand, it seems that this letterpress style commonly ignores, or only lightly applies overshoot, perhaps for historical reasons. I could just experiment and see what I like, but the final and most confusing part to me is how or if I should apply overshoots where there is a curve next to a serif like the "S" and the top of the "C" below.

Has anyone else explored these points before? Do you have examples of what you think worked well...or didn't? Thanks guys!

Nick Shinn

I think one has to approach this issue font by font, and trust one’s judgement by examining various settings.

For instance, the Thin weight of a typeface might require almost no overshoot, whereas the Fat might need lots.

Then again, one might vary the amount of overshoot between upper and lower case, even though the baseline alignment zone is the same for both.

Thomas Phinney

In this case, I think the vertical serifs (C, S) have large horizontal elements, and probably need to align with the flat tops. The rounds need some overshoot they aren't getting. But as your sample doesn't show flats next to rounds, none of this is very obvious or emphasized in the sample.

Ray Larabie

Don't measure or think about percentages. Increase the overshoot until it's noticeable, than back it off a bit. The round tops of the C and S require much less since the slab solved most the the problem that an overshoot would have solved. In cases like that you just need to add just enough overshoot so it doesn't look weird next to an O.

PabloImpallari

Choz, read http://www.frerejones.com/blog/typeface-mechanics-001/ and http://www.frerejones.com/blog/typeface-mechanics-002/

Choz Cunningham

Thomas, 
This is the handiest picture that has a few curves next to flats. Now I see it doesn't include a C or S. I'll make a sample with those mixed in shortly.

Ray,
If I don't measure, wouldn't it be too easy to get inconsistent overshoots from one letter to another - or is that acceptable? Or, do you mean don't measure the first one while I'm doing the tweak, and then measure it to copy to the others? I'll experiment with the C and S (and 9,etc.) probably last. 

Pablo
Wow, I never noticed there was a #2 to the mechanics before. As you can see fixing the contrast for optical effects is not something I've started. I try to stay as on-grid as possible until I've designed all the glyphs because I'll often be cutting pieces of one to make another. However, that is really going to help!

Nick, I'm not sure yet if I'm doing a lowercase, or a unicase.


(yes, I know the M is awful   )

[Deleted User]

The user and all related content has been deleted.

Max Phillips

Type of curve makes a difference, too. I think squarish rounds need less overshoot than more circular ones.

Choz Cunningham

I left out the S, but the G has an identical top.

Ray Larabie

If I don't measure, wouldn't it be too easy to get inconsistent overshoots from one letter to another - or is that acceptable?

It's acceptable. There are obvious ones where you'd want them to match, like the top of an O or the top of a Q, obviously. But the top of the G with the serif; there's no reason it would have to match the overshoot of the top of the O.

You'll encounter this question when you make a lowercase n or h. Sometimes it's the same overshoot as the lowercase o, sometimes not. You just have to walk around the room, get a coffee, come back, look at it again. Does it look too small? Too big? Observing optical illusions is tricky. It helps to walk away and come back to it with fresh eyes.

PabloImpallari

Set up your Blue Zones, and test, and decide at what size the overshoots should become visible.
You probably want your overshoots to stay invisible when the font is rendered from 10 to 20px, and only become visible when rendered bigger than 20px. (It can be 20px, or 30px, or 40px... or whatever size you decide).

Choz Cunningham

I ended up with much less overshoot than I imagined having. Less than 2%. I am assuming that Lager will be usually be used for large text. I find the overshoots of some of the nicest typefaces become distracting at super jumbo sizes.

Even this amount looks strange when I think about it, but seems fine when I'm looking at the letters and not the contours.





Ray, I did exactly the way you said, and it worked great. I also took a few breaks and tried lots of pairings, too. I ended up doing something like half as much movement on some of the punctuation, too. And I was surprised, but I ended returning to no adjustment on the S, G, and similarly serifed letters. No amount looked right with the slabbiness.




Thanks for the help, everyone.

joeclark

I was going to point you to the ancient article in U&lc that described the process of creating Lubalin Graph Condensed, but – thank you, Monotype – no links work. I thought it relevant but can’t prove it because the things won’t load.

Choz Cunningham

I just read a little bit on Devroye's page about Lubalin Graph. Which I love (uncondensed). I believe you would have been right, that would have been a good read. I may be able to find it digging in the fonts.com pdfs...? My current challenge is now trying to find a common metric for the sans and slab that doesn't sell one style short. The article might help with that too.

Stephen Coles

Joe, you’re right that the uandlc.com server is pretty unreliable. I often have to attempt a download several times. Usually I keep them offline for later reference, but it’ll be a while before I have all the issues. 

Using Google’s site search you can find all the references to Lubalin Graph in the U&lc PDFs. Unfortunately, I couldn’t find the example Joe refers to.