Designing for the Eye Optical Corrections in Architecture and Typography By Niko Kitsakis, June 2025 This article highlights a special aspect of both visual design and architecture: Optical cor­rections (or optical ad­just­ments, if you prefer). If you found my piece about legible typefaces interesting, you’ll likely enjoy this one as well. Since I included many visual examples that are size-sensitive, make sure you read this on a big screen and not your mobile device. Optical Illusions The Müller-Lyer illusion, pictured below, makes you think that, of two lines, one is longer than the other when in fact they are of the same length. It appears in virtually every introductory book on graphic de­sign and, of course, in books on per­ception and psy­cho­lo­gy. You might not have known it by name, but you must have seen it before: ← → Is the blue line on top longer than the one at the bottom? Click the arrows to find out. It used to be said that, in order to see the illu­sion, you must have grown up in an environ­ment where lines very similar to these would explicitly appear in your everyday surroundings – in architecture and furniture, for example, or when you had played with building blocks as a kid and so on. This is, incidentally, what I too was taught at my school of visual design some decades ago. However, I always had a hard time be­liev­ing it, and according to some new research, this is indeed not the case: Not only can the illu­sion be per­ceived by anyone, no matter their cultural background, it even extends (to some degree) to the animal kingdom and can be ex­pe­ri­enced in a haptic version as well. The reason why I never really believed that the Müller-Lyer illusion (and others like it) would work only on people who grew up in our Western culture is that, in the first place, it seems visually far too uni­ver­sal to begin with. After all, they’re just lines. You can draw them in the sand, and they can appear naturally in many circum­stan­ces. When I look at another person, my brain will assign certain values to the length of that person’s arms, legs, and torso, and memorise them as some variety of line as well. If the brain didn’t do that, we would not be able to tell if a body or facial features were out of proportion. So in my view, something dis­cern­ible only by people from a specific culture would have to be far more elaborate than an optical illusion of just a few lines. While that was only my personal gut-feeling, I soon dis­co­vered that other people had made similar obser­va­tions. One such time was when I first read Paul Renner’s 1940 book Die Kunst der Typo­grafie (The Art of Typo­graphy). The book covers pretty much every­thing you need to know to become a typo­graph­er, in­clu­ding that you should not rely on mea­su­ring things but rather on what your own eyes tell you: Der glaube an das Zählen und Messen verführt in allen Künsten zu den gröbsten Fehlern. The belief in counting and measuring leads to the grossest errors in all the arts. Renner happens to show the Müller-Lyer illusion in his book and writes about what to make of it and other visual illusions like it. Take a look at this book spread first: If you read books, you have an advantage. If you look closely, you can see a version of the Müller-Lyer illusion in the top right box of the left page. Renner describes every optical illu­sion on that page separately, but what’s interesting is how he con­clu­des that section. He says that the following applies to all of the optical illusions: Für die Anschauungsweise des Künstlers, des Schriftschneiders und des Typographen gibt es hier keine optischen Täuschungen, sie rechnen von vornherein mit den scheinbaren Formen; was scheinbar falsch ist, ist für sie falsch; und was scheinbar richtig ist, ist für sie richtig. Es handelt sich also nicht darum, ein genaues Zählen, Messen und Bauen durch ein unbestimmtes, schwankendes, willkürliches Gefühl, das niemand nach­prüfen kann, zu ersetzen, sondern man soll einfach den wirklichen Augenschein, dessen Feststellung ein sehr geübtes Auge erfordert, als maßgeblich anerkennen; er ist zwar „subjektiv“, aber er ist, wie die Gelehrten sagen, „intersubjektiv gleich“, also bei allen Menschen den gleichen Täuschungen unterworfen. Da alle bildende Kunst für den Augenschein arbeitet, so müssen wir ihm überall dort recht geben, wo er mit der Zahl und der Messung in Widerspruch gerät. From the perspective of the artist, the punchcutter, and the typographer, there are no optical illusions here – they assume from the outset the apparent forms; what appears false is false for them, and what appears correct is correct for them. It is therefore not a matter of replacing precise counting, measuring, and constructing with a vague, fluctuating, arbitrary feeling that no one can verify, but rather of simply recognising the actual visual impression – whose assessment requires a very well-trained eye – as autho­ri­tative. This visual impression may be “sub­jec­tive”, but it is, as scholars say, “inter­subject­ively the same”, that is, subject to the same (perceptual) biases in all people. Since all visual art works in service of what the eye sees, we must acknowledge the eye’s authority wherever it stands in contradiction to number and measurement. The most interesting part here is, of course, where he talks about the visual impressions of the optical illusions being “subject to the same biases in all people”. Renner recognised (and he was far from being the first), in other words, in 1940 what other people seemingly are only finding out now: That optical illusions affect everybody in the same way. But why would people in the visual arts or architecture concern them­selves with optical illusions in the first place? If you draw a circle, it’s just a circle, and if you build a house, you build a box – isn’t that all there is to it? You would be forgiven to think that way, considering the soul­less con­crete blocks by third-rate “architects” that have popped up all around us – those suicide-inducing, bleak and ugly manifestations of the Microsoft Excel mindset. But you would also still be wrong. There is – or at least there should be – more to architecture and the visual arts than that. A Perfect Circle People like Paul Renner, who design and draw typefaces, have always known about the intri­cacies of human perception and how to work in service of it. To illustrate all of this, let’s look at Renner’s 1927 typeface Futura. Futura is what is known as a geometric sans-serif – a type­face that looks as if it was drawn with a ruler and com­pass. It is made out of straight lines, sharp corners, and seemingly perfect cir­cles and semi-circles. Take a look at the following image, which com­pares Fu­tura to FF Bau, a sans-serif type­face that is not based on such strict geometry: ← → Futura on top and FF Bau at the bottom. Click through the images to get a feel for the geometry of the two typefaces. Given that Futura is a geometric typeface, why would it be important to know about optical illusions and the tricks that our perception can play on us? After all, it is constructed exclusively out of basic geo­met­ric shapes. Surely, optical illusions don’t factor into those. Unfortunately, things aren’t quite that straight­forward. Futura – like every other type­face – simply wouldn’t look as good if it wasn’t com­pen­sat­ing for the biases of our perception. Just because it is based more strictly on geometric primitives than other typefaces doesn’t change the fact that it is also still subject to our biased perception. Consider the image of the ring below, where the black line is the same thick­ness all around and the overall shape is a circle that is geo­met­ri­cal­ly constructed and perfectly round. If you really look at it, you will notice that it looks just a bit egg-shaped – that is to say, it looks taller than it is wide. Also, if you think of it as a watch face, the more hori­zon­tal parts of the line at 12 and 6 o'clock seem just a tad thicker than the vertical parts at 3 and 9 o'clock. Can you see it? Admit­tedly, it’s subtle, but it’s there. Click through the gallery and con­tinue reading: ← → 1) A geometrically perfect ring. 2) The capital letter “O” from a bold version of Futura. 3) The bright blue is the perfect ring from 1 while the dark overlay is the O from Futura. 4) The same as 3 but in outlines. Starting with the geometrically perfect ring in the first picture, Renner compensated for our perceptual biases by making slight ad­just­ments to the thickness of the line and the overall shape to arrive at the letter O for his Futura. The shape is now a bit wider than it is tall, and the hori­zon­tal parts of the line are slimmer than the vertical ones. I used a bold version of the typeface for this example, since here the effects of the adjustments are more easily visible. The ad­just­ments were applied to all weights of the type­face, of course – from light to extra bold – but are rather more subtle in the thinner versions. ← → The capital O from Futura in different weights. The second one from the right is the same you saw just before. Click the arrows to reveal the proportions between width and height for both the outer shape and the inner counter of each letter. I used an arbitrary value of 1000 as a reference width for both shapes . I hope you have noticed how in the series of letters above, the heaviest one is a bit more than just optically corrected. You can actually see that its form is not purely symmetrical anymore. The counter (the hole in the middle) is now notice­ably taller than wide, while the outer shape is noticeably wider than tall. The interpretations as to why this is will differ, depending on whom you ask. Personally, I think this reflects a com­bi­nation of factors: On the one hand, as the typeface gets fatter, it becomes more prone to exagge­rations in its proportions. On the other, it would look lifeless if it looked geometrically perfect under all cir­cum­stances . We will get back to this problem of seemingly perfect geometric shapes looking “alive” or “dead” later. Be aware that most of the examples here are shown at an ex­ag­ger­at­ed scale to make the effects more apparent. In practice, however, type­faces are rarely designed for such large sizes. For that reason, the following example includes a second version which – while still rela­tive­ly big – is closer to the size these optical ad­just­ments were orig­i­nal­ly intended to address (assuming that you are reading this on a large screen). I hope you will agree that now, the ad­just­ments make even more sense. This being said, let’s look at the capital letter G from Futura, which is, of course, based on a circle shape: ← → 1) The capital G as it appears in Futura. 2) The capital O from Futura is shown behind the G in blue to demonstrate the difference in geometry. 3) The capital G is adjusted to match the geometry of the capital O. 4) The G now looks comically out of shape, as if it had Popeye’s chin. The considerations that went into making this are really quite asto­ni­shing, aren’t they? Note also what happens when you look at the large G in the fourth image for too long and then switch back to the first. Now it almost seems as if the “chin” of the G intrudes too far into the imaginary circle . While this effect passes, it is a good demonstration of how easy it is to fool yourself when designing these letterforms. When you do this sort of work, you have to get up from your desk quite often and reset your eyes. This next example uses a different typeface called FF Tisa. The optical corrections that I want to demonstrate in this par­tic­u­lar example exist in all typefaces – including Futura – but they’re more no­tice­able in this one. Click through the images and see what happens: ← → 1) The letters M, O, E, and A as they appear in FF Tisa. 2) Guides are attached to the upper and lower horizontal strokes of the letter E. Note how the O overshoots the guides at both the top and the bottom, extending just beyond them. 3) The O is scaled down so that it just touches the guides. 4) The O now looks too small compared to the other letters. Once again, I’m using the capital letter O to demonstrate one of the many opti­cal corrections that go into a type­face. This is because the letter is based on a round shape instead of a rect­an­gu­lar one: Since letters don’t appear in isolation but next to one another, they have to work together visually. If every letter were based on a rect­an­gu­lar shape, like an M or N, this would be pretty easy. But round and di­ag­o­nal shapes, as in O or V, leave more unused space around them, which, in turn, can give a certain visual im­bal­ance to a string of char­ac­ters. The optical corrections applied by a typeface designer are meant to counteract those and other im­bal­ances as best as possible . These are but a few of the examples I could have given. Every letter, every numeral, even the dot on a lowercase i is drawn with these op­ti­cal principles and human perception in mind. I hope that you can now ap­pre­ci­ate a bit more the care and craft involved in making a typeface appear visually correct – which is a different way of saying beautiful. Architecture At the beginning of this piece, I asked why people in the visual arts or architecture concern them­selves with optical illusions (and therefore optical corrections) in the first place. I used typography to demon­strate why they do it in the visual arts. Let’s look at architecture now. You might think it far-fetched to draw a con­nec­tion between ty­pog­ra­phy and architecture, but the two fields are actually more similar than you might have expected. The typesetter or ty­pog­ra­pher has to plan a layout and a structure in which ele­ments like paragraphs, pictures, and head­ings fall into place. Similar to the dif­fer­ent parts of a building, these elements depend on one another to work both aesthetically and tech­ni­cal­ly. And just as a building’s foundation defines the constraints for what can be built above it, the choice of typeface – and how it is used at its small­est size – usually defines the layout grid, and thus how the rest of a book will be typeset . In addition to these similarities, both type design and architecture also passed through a lot of different stylistic eras together. Consider the two examples here: ← → 1) The most striking features of Gothic architecture are the narrow, vertical pro­por­tions and the pointed arches. Gothic-era writing mirrors these concepts. 2) In Jugend­stil architecture, everything seems floral, overgrown even. Windows are often oval or arched. Here too, the typefaces of the time reflected these tendencies. Finally, the work of both the architect and the typographer demands a combination of technical expertise and artistic sensibility. Which leads us back to the optical corrections: Would you call knowledge of these phenomena technical or artistic knowledge? I’m not sure myself – and frankly, I don’t think it matters. The important thing is that you cannot do your work properly as an architect or typographer if you don’t know about them. Truth and Beauty Contemplate for a moment how much time 2500 years really would be if you were to go back in time: World War II ended 80 years ago – you have likely not lived through it, cer­tain­ly not as an adult. The In­dus­trial Revolution started around 260 years ago, and Christopher Colum­bus reached the Americas 533 years ago. That may seem like a long time, and yet we have only gone back about one-fifth of the way we need in order to reach our goal: The year 432 BC – the height of Athe­ni­an culture in ancient Greece. This is the year in which construction of the Parthenon was fin­ished – the imposing and yet beautiful temple that dom­i­nates the Athe­ni­an Acro­polis to this day. The Parthenon as painted in 1871 by Frederic Edwin Church. I don’t know how many books have been written about just this one building, but they must fill a small library. The Parthenon is famous for its sculpture, its architecture, its political and religious significance, and its influence on Western art and design. However, what’s most important to understand about it is, in my view, that it stands as the antithesis of the Microsoft Excel mindset: Totally uncompromising in concept and ex­e­cu­tion, it is the polar opposite of mediocrity, a monu­mental middle finger to “good enough”. But why is that? The reason I asked you to imagine going back through time was so you could appreciate the following a bit more: Classical Greek temple-architecture of 2500 years ago sys­tem­at­i­cal­ly used optical corrections. While North­ern Europeans were still living in mud huts and fending off wolves and bears, the Greeks were building marble temples that com­pen­sat­ed for the peculiarities of human perception. The Parthenon is a temple of the Doric order. Think of it as an ar­chi­tec­tural style, one of three in ancient Greece. As is always the case with these things, the style evolved over time and became more sophisticated. I’m men­tion­ing this here because what is true for the Parthenon is also true for many other Doric temples. The Parthenon, however, is con­sid­ered the pinnacle of that development. It would be too much to go through all the optical corrections that went into the Parthenon, so I will concentrate on a few examples that should give you a good idea of how far the Greeks went in their quest for beauty. Below you see a front elevation of the Parthe­non . The temple has eight columns on its short and 17 columns on its long sides. Pay close attention to how nicely symmetrical everything is in the first picture below, and then compare it to the second picture: ← → 1) The elevation of the Parthenon drawn in perfect symmetry. 2) The elevation drawn as it was actually built. Before, I talked about how perfectly geo­met­ric forms look dead. You will agree, I hope, that this is exactly what is hap­pen­ing in the first picture above. Technically, there is nothing wrong with what you see in that picture. As a matter of fact, this is how any architect today would build such a temple (or more likely a classical façade for a different kind of building). This wasn’t good enough for the ancient Greeks, however. Not only does it look dead and sterile, but the columns and thus the whole roof section also seem a bit heavy at the top. Compare the first and second picture again with this in mind. You can see how the tape­ring of the columns in the second picture not only helps to elimi­nate the feeling of the build­ing becoming top-heavy, it now also seems to be more alive . This is interesting because if the Greeks had strictly adhered to mak­ing adjustments that were only corrections of optical effects, one should not be able to see those corrections at all. But in some places, they deliberately overcorrected to achieve precisely this effect of their structures coming to life. In the three pictures of the columns below, I ex­ag­ger­at­ed the proportions (by making the columns too short) to make the effect more notice­able: ← → 1) Doric columns drawn as perfect cylinders. 2) The columns are now tapered, the base being the same width as before. 3) The columns with entasis. The convex shape or swelling that the columns have in the third picture above is called entasis. It is to be distinguished from simple tapering as seen in the second picture. Scientific papers have been written about what purpose the entasis really has and whether it actually makes such a big difference visually. While I’m the first person to point out the im­por­tance of epis­te­mol­o­gy and the scientific method, I think this miss­es the point. Any person who is sophisticated enough ar­tis­ti­cal­ly will tell you immediately that this just “looks right”, compared to straight cylinders or simple tapering. Not everything can be counted or measured, re­mem­ber? No Right Angles The most famous optical correction in the Doric temple, however, is the curvature of the stylobate (the foundation on which the columns stand ) – and, by extension, of the entire build­ing. If you come from engineering or modern architecture, the idea of a curved foundation must sound like a sick joke to you: The whole base is slightly arched upwards in a convex shape. In the case of the Par­the­non, this means that along the long side of almost 70 me­tres, the stylobate rises about 110 millimetres, with a proportionate rise along the short sides. The long side of the stylobate of the Parthenon as seen ca. 1910 at the north-east corner. The angle of view makes the curvature more apparent than it would normally be. The columns are not perfectly vertical either: Not only did they have to be ad­just­ed for the slight curvature of the stylobate that they stand on top of, they also tilt slightly inwards (called their inclination). They are a bit more than 10 metres high but would meet at approximately 5 kilo­me­tres if they were extended. Finally, the entab­la­ture – which is ba­si­cal­ly every­thing that sits on top of the columns – is arched upwards like the stylo­bate. Geometrically, then, the whole structure is not simply a rectangular box, but best thought of as a truncated pyramid. To put it another way, these refinements mean that not a single right angle exists anywhere in the structure! Why do all of this? Because otherwise, the base of the building would look like it was sagging, and the columns would look like they were about to fall outwards. Being a typographer and visual designer, I think I under­stand the con­sid­er­a­tions of the ancient Greeks perfectly well. At the same time, they were also prag­ma­tists, and a convex floor never needs cleaning from rain­water… although this might just as well have only been a welcome side effect. Below is a visualisation of the curvature and inclination: ← → 1) The elevation of the Parthenon as it appears to the eye. 2) An exaggeration of how the temple would appear without the optical corrections. 3) The optical cor­rec­tions in an exaggerated form. There are many more things that I could go on about regarding Greek temple architecture and the finer points of visual design. But I want to leave the reader with a few things to contemplate instead: We live in a time when the availability of knowl­edge gets confused with having knowl­edge. People seem to think that having the internet at their finger­tips means they no longer need to know anything themselves. But in order to understand things, you need a lot of knowledge readily available in your head. Only then can the mind make the connections between the different points of data and come to new insights. This cannot happen when that knowledge is external, in a book or on some Wikipedia page that you have to look up first. You might wonder what any of this has to do with optical corrections, visual design, or archi­tecture. The point is that I see more and more people who – I have no nice way of saying this – know next to nothing and have little interest in the world around them. That in itself, you could say, is their own business. But these people go to work and make decisions. They are asked for their opinion. They approve fund­ing for design and ar­chi­tec­tural projects. And while many of them hold im­pres­sive-sounding degrees, their actual knowl­edge is far too super­ficial. These people will never ap­pre­ci­ate the sub­tleties I’ve written about here. And yet, I put it to you: These sub­tleties are precisely what make us human. If you don’t believe that or don’t want to hear it from me, I suggest you take a look at what Steve Jobs once said in an interview about es­sen­tial­ly the same matter. This first snippet from a video interview resur­faces online every couple of years: The only problem with Micro­soft is they just have no taste. They have absolutely no taste. And what that means is – I don't mean that in a small way, I mean that in a big way – in the sense that they don't think of original ideas, and they don't bring much culture into their product. That remark is important in itself and deeper than the dismissive tone might lead you to believe. But Jobs continues by saying the following, which is the far more impor­tant part. The emphasis is mine: And so I guess I am saddened not by Micro­soft’s success. I have no problem with their success. They’ve earned their success – for the most part. I have a problem with the fact that they just make really third-rate products. Their products have no spirit to them. Their products have no, sort of, spirit of en­light­en­ment about them. They are very pedestrian. And the sad part is that most customers don’t have a lot of that spirit either. But the way that we’re going to ratchet up our species is to take the best and to spread it around to everybody so that everybody grows up with better things and starts to understand the subtlety of these better things. And Micro­soft’s just, you know, it’s McDonald’s. And with that, you now understand why the Greeks built their temples on hills: They weren’t fortifications, they were built there so the Greeks could show their best and “spread it around to everybody”. That is the real reason why someone would arch the foundation of a 70-metre-long building upwards by the width of a hand, and why someone else would take enough time and care to make the shapes of letters fit one an­oth­er visually in the most subtle ways. None of this is strictly necessary, of course. You can get your work done using Microsoft Windows, and you won’t (immediately) die from eating at McDonald’s. And you can do these things while every build­ing you see out your window is a rectangular concrete block that looks like the filing cabinet of the soulless bureaucrat who had it built – all while you read your news in Comic Sans. But that is not exactly who we, as human beings, should aspire to be. Because if we let medi­ocrity win, we all lose what makes life worth living: Beauty and our culture. The Jugendstil illustration comes from a book at the archive of the university of Bologna while the gothic illustration is from a book at the university of Heidelberg. Share this article on Twitter/X or Instagram!