Italics angle compensations

Raphaël Ronot

Hello everyone,

I'm having a hard time findind anything online regarding the angle-variations that some — most? — of the italics present (example below : Garamond).



So my question is: is there a rule or some research/guidance that could help understand how to calculate those angle variations ? If rule or calculation there is. 

I get that this is a way to compensate a kind of "spinning" effect — for instance the f could appear like rotating clockwise if not somehow backslanted — but I'd like to know more. 

Any help is more than welcome ! 
Thanks in advance and have a great day / evening depending on where you're reading this from !

Raphaël

Ori Ben-Dor

Hi Raphaël,

It took me a long time, but I've learned not to look for such rules. You won't find any.

It's all about optical illusions and correcting them. These things are subjective and sensitive to details that vary from one design to another, so you'll just have to trust your eyes.

The idea is to make all letters appear as if they slant at the same angle. If all "vertical" strokes actually slanted at the same angle, mathematically speaking, we would perceive it as different angles. Those dominant terminals (serifs?) in /i, for example, pull it back and give the impression it slants less than it actually does. This is corrected by making it slant more. On the other hand, /l is less affected by its terminals: the top terminal has a different shape and doesn't pull the whole letter back as much, and the "vertical" stroke is longer and thus more dominant. So /l doesn't require the same amount of compensation as /i requires, though it probably still requires a little compensation, because of the bottom terminal.

You'll just have to play with the angles and pick for each letter the angle that makes it look right.

Cory Maylett

These kinds of things are more about art, aesthetics and intuition than they are about finding rules and formulas to follow.

Raphaël Ronot

Thank you @Ori Ben-Dor and @Cory Maylett 
I kind of feared that this was more a feeling question than a theory one, but your comments are helpful. 

If it's of any help to other people interested in the topic, I found this picture (taken from an article about the typeface Sindelar) giving hints about how and why corrections can be applied to specific letters. I doubt this can be made a rule of thumb, but it's a start. 



Until next time friends
Raphaël 

Paul Miller

As with most things in font design, adjust it until it looks right.  If you adjust until the theory/numbers are correct but it doesn't look right then it's wrong.  If it looks right but doesn't match the rules/theory it's right.

Your eyes are the ultimate arbiter of what is correct.

Vasil Stanev

It's all about optical illusions. I think the funny pages should be at the beginning of every course about design and art. "Which of these things doe not fit", "Find the 9 differences between the two pictures" and so on. Great, great training, IMO.

The right terminal of the h leads the eye to the right and with this the stem appears to skew rightwards. This however is countered in the d with the belly, so it does not need a compensation there. And so on.

Much can be learned about this when understanding a Greek temple. For example, why the columns lean inwards and taper off, and why they have a thing called an entasis at a certain point in their shaft . 

I also looked for hard rules at the beginning of my career, but with time the eye became the dictator. The best advice I can give it is to train your eye with practice and quantity.

Raphaël Ronot

Thanks a lot @Vasil Stanev @Paul Miller
At least now I have the confirmation that there's no rule haha !

notdef

> For example, why the columns lean inwards and taper off, and why they have a thing called an entasis at a certain point in their shaft. 

It is my understanding that the tapering is done to give the appearance of a flat façade when viewed from a certain distance at eye height, i.e. counteracting perspective distortion. (I think Gombrich discusses this.)

The entasis appears to be a different thing, and more relevant to the topic.