Getting a Basis for Specific Optical sizes

Jacob Casal

I occurred to me in my font to be designed for specific optical sizes that I was focusing too much on one size, so I’m trying to get a good sense of balance for the other sizes now before getting too far. This post is more for advice on getting the font’s optical sizes right early as the font’s not ready for fuller critique yet, perhaps some words of warning on things to avoid when dealing with optical sizes. Didots are often suggested for display type, but I want mine to be viable in running text as it was used in manuscripts and decently printable on the average printer at each size. I am using the /l to gauge the main values to use.

The serif height currently matches the spread of the ink seen in the manuscripts, I am designing for 8, 12, 16, 24, 42, 64, and 96 points. I made the serifs a little longer at 8 points, and will probably do the same for 12 and 16 points. (I liked HTF Didot’s size plan, 16 points even matches the manuscript I’m using, but the smallest sizes were a bit too small for the font.)

The attached PDF has vectors at their respective sizes on the first page, and a scan of the printed result on the second (the printed result is cleaner in real life, but there is a sort of thin, gray, subtle “drop shadow” appearance immediately under serifs). I appreciate you.

Jasper de Waard

First, I don't think you can estimate these things from a single l. You need running text at the intended sizes. Second: optical sizes are about more than just contrast, another reason a lone l does not suffice. Rhythm, spacing, kerning, ink management, proportions, the list goes on... and all these things interact.

Jacob Casal

Ah, your right, and I knew of that much at the very least, in my naïveté I oversimplified things. I’ll exercise more patience and do it right!

Claudio Piccinini

Even when dealing with a new design, the best thing is to look at actually printed samples in actual point sizes (keeping into account that ink will have spread a lot more in the smaller sizes). And then apply the considerations to your new design.

Here’s a preliminary evaluation I was doing on some glyphs with Howland, also to consider width consistency. Right to left these are sizes ranging from 12 pt to 60 pt (text sizes can even be more sturdy, but I’d avoid an almost monotonal treatment if the design has contrast to begin with).

Claudio Piccinini

P.S. Note also the changing position of the middle horizontal stem across weights.

Jacob Casal

Thank you for the advice Claudio, thanks as well for the example! There’s definitely a lot for me to cover here so knowing about a few of the subtleties to consider is always a great help.

Jasper de Waard

Why do you even want optical sizes to begin with? If I were you, I would just start with designing a very nice text OR (exclusive or!) display face. Once that works, you can always try to translate it to different optical sizes. Your ambition is admirable, but take it step by step.

Jacob Casal

Oh… I like the challenge of it, and like a full immersion language learning experience one has the opportunity to learn some nitty-gritty details of the tool and how to use it. But, even then you start simple, so yes, maybe focusing on the 16pt the manuscripts are actually set in is a good place to figure out the basics and make the design solid instead of doing everything in concert in relative confusion. I’ve learned a lot so I should apply that first as a sort of refining test before getting into the big stuff.

There is more to my wanting it optical than just the challenge though, a linguistic reason. I do like Didot, and grown to like it even more as I’ve researched the works, but it’s more of a co-main reason I chose to design this particular version, but I’ll get into that when the basics are done first.

Claudio Piccinini

Jasper de Waard said:

Why do you even want optical sizes to begin with? If I were you, I would just start with designing a very nice text OR (exclusive or!) display face. Once that works, you can always try to translate it to different optical sizes. Your ambition is admirable, but take it step by step.

I believe the problem lies mostly in the fact that Jacob wished to tackle the things all at once: if you start reasoning across sizes keeping simple considerations under scrutiny, it can be an effective educational exercise. You can reason in terms of both text and display, but of course having previous experience (at least a bit) with both  helps and, as you say, the goal must be taken in small sizes, step by step.

Claudio Piccinini

P.S. A Didone/Bodoni is also probably not the best genre to use to learn, IMO.
It is elegant but also is quite rigid. Not by chance, ITC when did the new Bodoni with optical sizes in the 1990s, took hints from the original lead types (even travelling to Parma, here!) and adopted quite a different approach for – say – 12 and 72 point sizes.

Claudio Piccinini

See what I mean…

ITC Bodoni Seventy-Two:



ITC Bodoni Six (Ugh!)

Claudio Piccinini

Jacob Casal

As I said, I do like a challenge. Joking aside, I do hear you, believe me. When I did a little bit of work on the 8pt /l it felt odd for it to be so chunky. “Ugh!” indeed (for it and the Bodoni), but I reminded myself that (hopefully) it would be used at the intended small size. Our chunky ITC Bodoni 6 at 6 via the website:

But of course that’s been refined to a tee to work by its designers, and I’m just starting out and need to take it slow. Oh yes, when it comes to optical sizes Didot is probably not the best place to learn it through due to its rigidity in such matters. My next project would take a step into more humanist type (of my own design). However, I’m still set on Didot right now, having turned everything down to a narrower scope to get it right: line upon line, tittle after tittle and all that. Still the Didot? As I said, there is more to my choice here than just Didot’s Latin and Greek , but I’m not there yet regardless.

I appreciate all of your voices of concern though! It helped me take a needed step back that I was too stubborn to do myself.

Craig Eliason

I think perhaps the rigidity of a didone will help make stark the design problems. In that sense it may result in perhaps not a better font but a better learning experience. 

Claudio Piccinini

Craig Eliason said:

I think perhaps the rigidity of a didone will help make stark the design problems. In that sense it may result in perhaps not a better font but a better learning experience. 

Possibly. Yes, the outcome might not be satisfying at first, but of course you can learn regardless. My impression is that the verticality of the stem widths ratio makes modern romans more an exception than a rule. I would experiment more with an existing display design, looking also at lead examples.

Kent Lew

There is, of course, modern precedent in HTF Didot, which was published in seven different optical sizes. Analysis of Jonathan’s choices there may add to the learning experience.

Claudio Piccinini

Kent Lew said:

There is, of course, modern precedent in HTF Didot, which was published in seven different optical sizes. Analysis of Jonathan’s choices there may add to the learning experience.

What I find curious in Hoefler's approach is that he chose to keep the x-height the same, even in the smaller sizes. The /D/ is almost narrower in the 6pt master, I wonder why.

Jacob Casal

Oh yes, I had been looking quite a bit at Hoefler’s work, the Try page was a bit of a mess for me, but nothing Inspect Element couldn’t revise.

John Savard

ITC Bodoni Six? Nothing wrong with it that renaming it "Clarendon" wouldn't cure.

Jacob Casal

It is a little funny that a modern, thin-serif typeface becomes more of a slab for its smaller sizes. I wonder what would happen if someone took something like Egyptian Slate, made it the smallest size and optically sized it for larger sizes, but now I’m digressing even more.

Claudio Piccinini

John Savard said:

ITC Bodoni Six? Nothing wrong with it that renaming it "Clarendon" wouldn't cure.

Not at all. It’s closely based (in fact, even too closely, as it preserves print imperfections from the books) on the same point sizes of Bodoni faces they gathered in Parma. If it was cleaned up a bit it would be better, IMO, but I liked the rationale behind their approach.
And the fact these digital versions were meant for print.

Claudio Piccinini

Jacob Casal said:

It is a little funny that a modern, thin-serif typeface becomes more of a slab for its smaller sizes. I wonder what would happen if someone took something like Egyptian Slate, made it the smallest size and optically sized it for larger sizes, but now I’m digressing even more.

You can keep the contrast, to a desired degree. The smallest point sizes Bodoni used kept a certain contrast. Obviously inking and printing decrease it, but that’s how you actually experienced the typeface in books.
Bodoni-inspired later faces used since the post-war, when printing quality improved, when used in books for long passages of text, quite always gave me the headache. High contrast is never good for small size text, especially in essays and books.

Jacob Casal

Headaches in print and “dazzling” on screen, mm definitely things I have been considering, but I’ll figure out something! That’s probably why it’s hard to find small text used across the manuscripts. In the ones I’m using, there are four sizes used (speaking in regular points and not Didot points): 16pt for the main body of the manuscripts, ~18pt for a short introductory section (40 pages and 20 pages in two separate volumes) and ~14pt for the footnotes that only appear in those introductory sections. There is only one instance of five words in ~11-12pts used for a scholar’s credentials. It’s a bit hard to gauge what’s going on with the italic sizes though. I’m not sure what’s going on there, even when accounting for the curve of the page.