In a grotesque typeface some of the capital letters seem either to narrow or too wide. How can they be corrected?

Nymus's picture

In a grotesque typeface some of the capital letters seem either to narrow or too wide. Traditionally, the capital letters are categorized width-wise as follows:

Thin I J
Medium HATVX
Ultra Wide MW

Is there any method or procedure to correct the ‘wrong’ capital widths?

Rob O. Font's picture

>Is there any method or procedure to correct the ‘wrong’ capital widths?

Type design.

Cristobal Henestrosa's picture

I assume that you mean by “correct the ‘wrong’ capital widths” to design capital letters with classic proportions (i. e., something in the Futura or Gill Sans style). I don’t know if a grotesque can benefit from it, but I can tell you AFAIK there is no method of doing this, except trying, as David said.

In fact, IMHO those modern proportions are an important part of a grotesque, but I can change my mind if you demonstrate the opposite. Good luck!

hrant's picture

Thomas, all caveats concerning legality/ethics and stylistic appropriateness
notwithstanding, to elaborate :-) on what David wrote: if you decide to get
a third party to do such modifications for you, you might want to ask for
quotes from type designers. My email address: hpapazian at gmail dot com.
There are of course quite a few others here who could carry out this sort of
thing very capably.


Bert Vanderveen's picture

Some of these ‘mistakes’ are the result of technical limitations, eg the mix of positions in a typesetting matrix (where some characters had to conform at the cost of proportions) and the unit system used. I am talking about the olden days aof course, when typesetters received a quota of milk to drink during their work day, to offset the effects of lead fumes.

There is a certain charm, I feel, in these ‘compromised’ typefaces. Why change that? Just use a modern face like National.

. . .
Bert Vanderveen BNO

Nymus's picture

Thank you all very much for your input. You have raised quite a number of interesting issues to be tried out. While waiting for your replies I have thought about it a little more and came up with the following alternate approaches to this particular problem:
1. Use Type Design like David said.
2. Using an empirical method, i.e. some form of numerical analysis as follows: Select a number of Grotesque typefaces such as Akzidenz (A), Franklin Gothic (FG), Helvetica (H), News Gothic (NG)… with the same weight as the ‘grotesque’ [actually, lets say that it is a new Grotesque typeface under construction by a novice typeface designer] with the probably erroneous capital widths.
For each grotesques selected take the widths of all capitals and divide them by the width of ‘H’ e.g.
News Gothic:
A/H (), B/H (), C/H (), D/H ()…
Do this for the other selected Grotesques.
Do this also with the Grotesque which contains number of erroneous Capital widths (GECW).
You end up with 4 series. Average them. The result would be a generalized empirical formula or Capital widths relationships.
Compare each element of the Averaged capital width ratio series (ACW) with each respective element in the GECW.
At this point the Capitals with the erroneous widths will stand out by virtue of the fact that they would have the largest discrepancies.
3. Another method, an optical one would be to print out lines all the capitals of the selected Grotesques and lastly of the GECW. Take your time looking at them. The Capitals with the erroneous widths in GECW will stand out.

hrant's picture

> Using an empirical method

Although I'm partial to this sort of thing, and the intentions are golden, there's a severe limit to how much this is worth the effort. It's great for understanding fonts, but lousy for actually producing fonts. The eye is much more effective in this regard. The problem is, without experience and/or quality feedback, even the eye is not enough. If you still want to attempt numeric analysis, try to find a copy of the fascinating Unica booklet by Gürtler et alia. You might also want to study Nick Shinn's Richler, which IIRC exhibits an interesting "third way".

What I suggest is giving it a shot, posting
the results here, and hoping for the best.


Syndicate content Syndicate content