So it seems well established that a key part of being able to design type is the ability to see type; that is, to see all the optical illusions and being able to judge colour, optical spacing etc. But I was thinking, how do you keep the ability to see these things consistent and what if you begin to overcompensate?
For example when looking at overshoot, isn't it possible that a type designer is so used to seeing and being aware of overshoot that over time they require more and more overshoot for it to *look* right? Optical illusions are just that after all. And how would this play into catering for the everyday viewer reading type?
Another example, the theory that we see diagonal strokes—right down to left as thicker than left down to right, is super interesting if it is due to writing techniques. I wonder if this illusion is stronger for people who have experience in said writing techniques or if type design. Am I just overthinking things?
Comments
I have been thinking about this (!) from time to time, since then.
When I was teaching Schrift, I said to my students: link seeing to thinking! For type design (and other disciplines) it is essential to understand what you see. That may sound trivial but it isn’t.
When you learn to see – overall important, right! – that is not about illusions. It is about visual facts. You mention good examples: overshoot, thicknesses. When you have learned to watch these phenomena properly you won’t demand more of it over time, but you keep the ability to see and maintain the balance of things or effects, which, I would say, is crucial.
The only thing for me that is problematic is after long periods of working only on italic fonts, upright text starts to look strange. This does not last long so it is not difficult.
I think the key is being able to see things in the environment they will be used at or at least that they’re aimed at. So for instance, if you are deciding on the overshoot size for a display face, then you should really judge if your initial guess is correct at display sizes.
The optical compensations our eyes and brain do are size-dependent, so in my opinion it is hardly possible to make one set of outlines that will look perfectly compensated at all sizes. That can only be achieved with optical sizes—whether interpolatable or not.
A good rule of thumb is to start with uncompensated shapes: make the lines / angles / measurements, etc. actually geometrically correct / aligned / symmetric, etc.
Then take a step back and look and see what you feel: does something feel low or light or high or bold? You’ll be surprised when things that are mathematically aligned actually look misaligned! A typical case is the arms of an X. Once you have identified something that feels off, see if you have the same feeling when enlarging or scaling down the glyph. If so, make some adjustments and judge them at the size you care most about.
It is of course possible to start developing habits and thus overcompensate when there is no need, like you were mentioning. I find it’s good practice for me to ask colleagues if they see the same distortions as I do. In order to second-guess your hunch it also helps to tilt your head sideways, or to flip or rotate the shape to see if the distortions you see are emphasized or removed. (This is also a good way to see if a curve is smooth or not, by the way.)
Try to focus on one aspect at a time when doing compensations, to isolate it from other variables in your design. Sometimes compensations are related, for example on round shapes the amount of vertical overshoot at the top is often the same as at the bottom, but more often I think the compensations can be done independently. Once you apply them to one glyph then it’s worth going over other related glyphs and seeing if the same compensation is necessary there, too. I find it is generally the case, but you may also find exceptions this way.
Correct, evaluate & repeat.
As to your question on whether the corrections are necessary because of how our eyes are used to a specific writing system or not, consider the fact that what native Latin readers world consider “reverse” contrast is considered “normal” by the Arabic readers, and vice-versa. That said, I would also be wary of trusting anecdotes like that to try to explain visual phenomena. The processes that happen between the letters and our brains are very complex and still not fully understood.
I think it’s the same as musical ears – you learn how to hit the right note, not higher, not lower.
If you think you overcompensate things, just forget about compensations and put your typeface into context; you’ll see any mistakes if they are there.
One thing that we type designers tend to overlook is the size of the shapes. We focus too much on the design part while there is also the legal and the coding side. We work at great magnification (e.g. 700 pt) and tend to believe everybody sees type as we do, when in reality when we set the product in to the wild the type is so small (10-18 pt) that all the kinds of optical compensations get completely lost. It is much, much more important for the code to be right than for the design to be impeccable, same goes for licensing. I am always reminded of great actors or musicians that write a book or draw a painting. Did you know Viggo Mortensen (Aragorn) is a painter? He painting are less perfect than his acting (which I highly praise), but they got sold for a pretty hefty sum. Paul McCartney fancies himself a painter. He's not by any stretch, but made a buck or two from his work, as far as I gather. My point is , if Aragorn and McCartney can sell rubbish subperfect products, there is no reason for us to let an anchor point slip under the rader here and there if the type sells and deadlines are kept. Most people wouldn't know good from bad type if you educated them for hours.
To summarize: 1. get the basics right (which parts should be thicker then others, which thinner, don't set the type too tight, know these-and-these pitfalls, etc.) 2. polish your skills with many fonts and learn the details. The eye will get trained, no worries.
I often find when coming back to something I worked on for a while with fresh eyes, new issues arise and issues I thought I resolved are still there. I wonder, is it generally a bad idea to work solidly on one single issue?
I'm sure eye fatigue plays a role in all of this. I'm curious if anyone have any insight into general workflow in regards to keeping your perception "fresh". Do you make small incitements over lots of different glyphs over a period of time? How often do you switch between digital work and physical proofing? How often do you take breaks? etc.
This is a characteristic of many creative endeavors, and is usually considered to be a good thing.
For instance, when writing a text, a usual advice is to stop when you feel you are (nearly) done, and then look at it again after a while with fresh eyes—because then you may see a lot of things which can be improved.
For example, when the kerning is correct there’s still more space between AV or LT than between HH, which sometimes confuses me.
The simplest beginner mistake is just to kern too tightly in general, so that kerned pairs look tighter than combinations that didn’t need kerning in the first place. Not everyone makes this mistake, and if they do, they usually get over it pretty soon.
The next error is to kern the "weird combinations" too tightly. LV, LT, AV, WA, Vo, Wa, To etc. The general rule of thumb is that the more of the glyph lengths that are close to each other or intertwine, the the more you can back away from making them mathematically the same distance apart as, say, the simple verticals. So because "AV" has a longer distance close together, those two lines should not be as close as with "HH"; leastways, that’s how I think of it.
But how much looser? It depends on whether the typeface is optimized for text sizes, or very large sizes. I am still sometimes surprised by how loosely these latter combos are kerned in text typefaces by master type designers! In display (very large) sizes, they can be close to as tight as you might have been tempted to do. In text sizes, you might be kerning half that much. Or something like that.