Research about optical ajustments in printed small typefaces

afonseca1974's picture


I'm doing a research about optical ajustments in printed small typefaces.
I need to understand better what adjustments are made to Type Designs to be used in printed small typefaces.
I've managed to gather some material:

-Finer points in the spacing & arrangement of type
-Size-specific Adjustments to Type Designs: An Investigation of the Principles Guiding the Design of Optical Sizes
-The optical scale in typefounding
-Revisiting the Concept of "Typeface" and the Optical Scale in Typefounding
-From the Optical Scale to Optical Scaling

Doe anyone here can point more stuff? Articles, dissertations, books?
Thanks in advance.


charles ellertson's picture

I don't know if they've released much information, but Adobe has surely done work on this -- witness their optical sizes, and the newer "catalog" fonts.

And it depends somewhat if you're after theory -- academic study -- or results in the contemporary digital world. If the former, by all means include Dowding (and Factors in the Choice of Type Faces London: Wace, 1957, would likely be a better resource than Finer points...). But if not, contemporary studies should be more accurate in details.

William Berkson's picture

Slimbach's pamphlet Designing Multple Master Typefaces is still a very good source. Tim Ahrens has written a book about this, and it is being revised for a second edition. I haven't had a chance to read it, unfortunately.

Some optical adjustments such as larger x-heights and looser letter spacing for small sizes I think are pretty universal, and traditional. But there is still a lot of experimentation, with different designers trying to solve these problems each in their own way.

Theunis de Jong's picture

Donald Knuth address this very tangentially in his "TeXbook":

What's the difference between cmr5 at 10pt and the normal 10-point font, cmr10? Plenty; a well-designed font will be drawn differently at different point sizes, and the letters will often have different relative heights and widths, in order to enhance readability. (Donald E. Knuth, The TeXbook, 1996: 16)

and a few similar loose remarks. It's a standard feature of TeX, but probably only used with the original set of Computer Modern fonts, not with any modern (scalable) Type 1 and OTFs.

afonseca1974's picture

Thank you all for the input.
It's an academic study but with a pratical part.
I'm doing a revival of a typeface designed in the 80's for print for small size that was never finished and digitalized.
I want to understand better the adjustments made, like making the bowls of "b" and "d" small to give more room to their ascenders, etc...

charles ellertson
"Contemporary studies should be more accurate in details".
I do believe that they could put some input too. Do you know any?

Thanks again!

Michel Boyer's picture

There is a 2009 typophile thread related to that subject:

hrant's picture

Indeed the old folks didn't document technical stuff like optical scaling almost at all, so you won't find much pre-digital. Exceptions: Christian Paput's "The Punchcutting" (you can tell he's French :-) and "Counterpunch" by Fred Smeijers.

Here's another angle: carefully study fonts with superb optical scaling. Two greats:
- Linotype's metal Primer, which AFAIK was the height of optical scaling pre-digital. There are some Typophile threads that feature discussions on it.
- Carter's MS Sitka. See: &

making the bowls of "b" and "d" small to give more room to their ascenders, etc..

Wait, you found a font outside of Fleischmann's work that does that?! Show & Tell!

Michel: I hadn't seen that thread – thank you!


afonseca1974's picture


I think that, if not all, some typefaces developed by ladislas Mandel for phone directories have the bowls of "b" and "d" is a small sample of Nordica Medium

hrant's picture

Oh, Mandel. The Magician.
BTW you mean Nomina, right?


Nick Shinn's picture

Hobo has no descenders, and thus compact g, p, q and y, but for a different reason.

afonseca1974's picture

Nomina is a different typeface...I really mean Nordica.
Nomina was developed for the Italian telephone directories (SEAT). Commissioned in 1999 to Piero De Macchi, it was first produced in 2001 . It replaced Mandel's Galfra which had been in use since 1977.

Nordica its a typeface developed by Ladislas Mandel for phone directories in the 80's. Some more info in this article (in French).


hrant's picture

Ah! Those are keeper PDFs - thank you.


lindenhayn's picture

> It's a standard feature of TeX, but probably only used with
> the original set of Computer Modern fonts, not with any
> modern (scalable) Type 1 and OTFs.

the entire original range of grades was preserved in the Type1 and OTF versions:

14.08.2013 16:41 111.872 lmroman5-regular.otf
14.08.2013 16:41 112.212 lmroman6-regular.otf
14.08.2013 16:41 110.792 lmroman7-regular.otf
14.08.2013 16:41 111.948 lmroman8-regular.otf
14.08.2013 16:41 112.680 lmroman9-regular.otf
14.08.2013 16:41 111.536 lmroman10-regular.otf
14.08.2013 16:41 110.400 lmroman12-regular.otf
14.08.2013 16:41 110.248 lmroman17-regular.otf

And it's fully functional, too. Just like in the old days, TeX (+ fontspec) will automatically use the appropriate grade for \tiny, \footnotesize, \huge, etc. ...Provided you're bold, or masochistic, enough to use that typeface at all.

Michel Boyer's picture

The lm* fonts are latin modern. The Computer Modern (type1) fonts also come in those "grades"; they are the cm* fonts; here is the list for computer modern roman:


Small sizes are used to render indices and double indices in mathematical formulas.

Here x is cmmi12, the first index i is cmmi8 and the second is cmmi6.

nina's picture

You may also want to get in touch with Adam Katyi, who did some work on optical sizes for his TypeMedia graduation project (last year).

quadibloc's picture

Indeed the old folks didn't document technical stuff like optical scaling almost at all, so you won't find much pre-digital.

While that's certainly true in the sense in which you meant it - that they usually didn't write about how they did it - there is other data that they left behind besides the fonts themselves.

For example, in the case of Monotype typefaces, there would be the set widths for the different sizes of type. In the case of the ATF foundry type catalogs, alphabet widths are listed, the sequence of which shows the scaling they used for the widths of the different sizes. (And, thanks to Typographical Printing-Surfaces, plus mentions in the ATF catalogs of the widths of spaces, I know that their foundry type, regardless of point size, was made to a unit of 1/4 point - so unlike the case with Monotype, individual glyphs will have had different scale factors applied at any one size.)

Syndicate content Syndicate content