Question: Why are foot serifs cupped?

Karen Cheng

Hello! I am a bit of a lurker on TypeDrawers, but I thought I'd try to make my first post.
I have a question about foot serifs that are cupped, as in the sample image below.

I have read several different reasons:

1) In Sofie Beier's book "Type Tricks" (which I think is quite helpful) and in this article by Mark Jamra, both authors note that an optional illusion can occur where serifs appear to bend outward. This illusion is corrected by cupping the serifs. However, I have never really been able to see the optical illusion that is described—flat serifs look okay to me? Perhaps my visual acuity isn't good enough.

2) It seems clear that some typefaces are placed on calligraphic models where the original calligrapher finished his/her letters with cupped serifs. That is, they drew their serifs with a curve. In this case, cupped serifs are a stylistic feature...yes?

3) This is not as well documented (or perhaps I have not found the right source), but in letterpress printing, sometimes ink spread around the imprint of the type. Did this ink cause the cupped effect? And if so, did the original type designers and punchcutters want the cupped serifs to be seen in this way? I didn't know if in this case, cupping was used to counteract the ink spread, so that the serifs would look relatively straight.

In this last case, I thought the situation might be a bit like contemporary ink traps. 
That is, some designers like how the ink traps look in fonts like Bell Gothic, etc.
And eventually, what was a functional detail becomes a stylistic feature. 

At any rate, if anyone has thoughts or helpful references on this, I would be grateful for input and suggestions. 

John Hudson

4) It looks nice.

Simon Cozens

Fred Smeijer's theory, from Counterpunch:

Igor Freiberger

I share with Karen this question and I am grateful to her for bringing it for discussion. My impression is that cupped bases are an artifact residual from calligraphic heritage, but a better informed answer is very much needed.

Regarding Smeijer's explanation, I am not convinced. In 21.2 he says "one deviation and the whole ideal falls apart". Why and when the contour would get any deviation? Bad printing? And in 21.3 he says these not-so-rational forms works better in text, which is a statement without a rationale (in this fragment at least).

notdef

Are we conflating cupped serifs



and bracketed serifs




?

Cristóbal Henestrosa

Regarding Smeijer's explanation, I am not convinced. In 21.2 he says "one deviation and the whole ideal falls apart". Why and when the contour would get any deviation? Bad printing?

Hello, Igor. I think Smeijers is talking about how making punches in XVIth century in some aspects was not a very precise practice, specially if counterpunches were used, because it implied striking metal bars with hammers in real size letterforms, so it is was better for the punchcutter to construct forms that allowed some inconsistencies without looking extremely weird.

Craig Eliason

Simon Cozens

I think Smeijers’ point can apply more generally: curves give you visual margin for error, whereas discrepancies in straights are far more noticeable.

John Hudson

My take on this is that the shape and treatment of serifs should correspond visually and stylistically with the shapes and treatments of letters in general and, perhaps more specifically, with those letters in which serifs are not a feature such as a — in many types — e g etc., in order to tie the whole together. In an renaissance style such at the Garamont shown above, the quality of line, curve and terminal in those letters needs to be reflected in the treatment of serifs. If the serifs were too regular, flat or rationalised, they would be out-of-style.

In a 19th Century style such as Clarendon, with sharper terminals and more rationalised features, flat serifs are appropriate, and cupped serifs would look wrong.

The 18th Century Neo-Classical and Romantic styles are an interesting arena for varying approach to serif design, because one can either emphasise the rationalist aspect of these styles, as many modern versions of Bodoni and Didot do, and make the serifs very flat and monolinear, or one can reflect the traditions of writing masters and engravers that inspired those types and make them softer as in ITC Bodoni.

ivan louette

There are so many variants, and each one has its logic. Often Venetian revivals have cupped serifs. The most cupped one seems to be Schneidler. Otherwise there are also some asymetric and slanted cupped serifs like in Alegreya.

Karen Cheng

Ah, I have seen that diagram in Counterpunch (I love that book–my favorite part is when Fred begins trying to make his own punches, and he writes "my respect for the contents of the Plantin museum went up by the minute...")

However, I think that diagram is more about the curved counterspace rather than cupper serifs. In the drawing, both serifs have flat bottoms. 

Still, perhaps the same thinking applies—that it was difficult to make the serifs precisely flat, therefore they were made curved. And to take John Hudson's point, perhaps this simply looks better.

Maybe the cupped serif is due to all of these things rather than one single thing...

Robin Mientjes

To build on John’s point, if you draw a rectangle and bend the left and right sides concave, then the top and bottom will be affected optically – it could make the top and bottom look ‘too straight’ or even convex. At the very least, it would turn all the corners into more acute angles (less than 90°), and that in its own right affects the perception of the shape. So if all stems are treated as curved segments, it isn’t illogical to apply the same treatment to other straight or straight-seeming segments.

Craig Eliason

kcheng said:

Still, perhaps the same thinking applies—that it was difficult to make the serifs precisely flat, therefore they were made curved. And to take John Hudson's point, perhaps this simply looks better.

I’ve never cut a punch, but I would think short contours on the outside on the punch like foot serifs could be made flat quite easily with a file—maybe more easily than adding cupping. Developing Smeijers’ point it would make more sense to say that because precise flat lines are difficult to achieve all around and inside the letter, it might look incongruous to employ them only on the easy parts. 

But it I do think cupping reflects a visual intention more that a production artifact. 

Florian Hardwig

Nick Shinn

I tried a convex serif effect in Paradigm, but on a practical matter, overshoot in the capitals was problematic. Note “HO”—the convexity counteracts overshoot, and yet realistically in this “paradigm”, I felt that the O needed to be the same absolute height as vertical stem glyphs, so the result was that the round caps ended up looking a bit short. And there were other difficulties in drawing the capitals.

Mark Simonson

I think cupped serifs tend to make a face look a bit crisper or sharper. You can also see this effect with Optima, with not just the stroke endings, but in the strokes as well. The inverse treatment of Cooper's serifs gives it a soft appearance, underscoring this phenomenon. See also rounded strokes on sans serif faces.

ivan louette

@Mark Simonson I think the crisp effect is caused by the vibration of cupped serifs. In my font Geranium, which has no angles and only very short straight lines, I designed cupped serifs specially for drawing attention to the base line and the x height horizontal. I think cupped serifs are also playing like a kind of roots.

Peter Baker

ivan louette said:

I think cupped serifs are also playing like a kind of roots.

I was also thinking about roots, following this thread, because of the way the cupped serif extends a little below the baseline. In fact, the relationship of the outlines to the baseline becomes much more complicated in a font like Ivan's Geranium, as the outline only touches the baseline at the tangents formed by the curved bottoms of the serifs (and it doesn't have to touch it even there). The font no longer sits on the baseline at all, but plays around it, like the way a jazz performance plays around the underlying beat but never actually plays it.

John Hudson

Nick:

I felt that the O needed to be the same absolute height as vertical stem glyphs, so the result was that the round caps ended up looking a bit short

In retrospect of which, wasn't the feeling that they should be the same absolute height incorrect?

Nick Shinn

Yes, but so was the feeling I got when I gave the round caps a greater absolute height.

I suspect that the “undershoot” of cupped serifs harmonizes nicely with overshoot in general, and is part of the same phenomenon.

Notably, PostScript Type 1 recognized this by adjusting both in its method of Alignment Zones. I vaguely recall that there is a special category of Serif Hint, and you have to make the round apex BCP of the cup exactly in the middle for it to work properly. Is that right?

Peter Baker

Nick Shinn said:

I vaguely recall that there is a special category of Serif Hint, and you have to make the round apex BCP of the cup exactly in the middle for it to work properly. Is that right?

The flex hint, explained here:

https://www-cdf.fnal.gov/offline/PostScript/T1_SPEC.PDF 

on pp. 72-3. The round apex doesn't have to be exactly in the middle, but the other requirements are quite stringent. There are diagrams to help.

ivan louette

@Peter Baker Thanks for the Jazz ! I love it (I played clarinet a long time ago). Geranium is far from beeing perfect, that's a fact. But that was my first try to draw a roman text font… And I had already several challenges in my mind : 1° how to avoid coercitive and hypnotist reading (I told it already in another thread) and 2° how to draw a soft font while remaining just nervous enough. The first of these two constraints was met when I saw my first Jenson prints. What differenciates Jenson characters from Griffo's ones is their low contrast. For example the top curves of Jenson's /m and /n have at least the same width than the verticals. That, with eventually stronger serifs, leads to a lineage that I call, in my own simplified language, picture film fonts (where counterpunches are the pictures and the black parts are the black areas around), while Griffo's romans characters, with their high contrast and prevalence of verticals, lead to what I call (also in my simplified and naïve language) a lineage of barcode fonts. For some personal reasons (perhaps physiologic) barcode fonts (like Times New Roman for example) hurt my eyes and that's why I privilegiate Venetian ones in my work. Another interesting thing in Venetians is that they facilitate the reading of several lines togheter and they allow your eyes to walk more freely on the page. I feel their dynamics slightly closer to ancient cursive non gothic calligraphy, and cupped serifs are perhaps better adapted to them for this reason.

Karen Cheng

Thanks to everyone for your answers and comments! Very much appreciated!

Craig Eliason

Another potential factor not yet mentioned: 

As metal sorts are used over and over, their surfaces get worn, and more so at their corners. So cupping serifs may also be seen as a countermeasure to the eventual wear the sorts can expect. In a way, a kind of subtle "light trap" at the corners. Zapf once talked about how the tapering strokes of Optima were intended to resist the wear he saw compromising standard grotesques.

We may find it useful to flip the question: not "what was the purpose of the serifs getting lighter in the middle," but "what was the purpose of the ends of the serifs getting thicker?"

Hrant Հրանդ Փափազեան Papazian

Possibly repetitive, but:

1: It counteracts an optical illusion of bulging. But I'm not convinced.

2:

(Which personally makes me cringe.)


Ergo: Generally best to avoid it.