Designers are taught often from the first typography course they take that kerning is a crucial aspect of typography. Type not kerned is a sign of designer sloppiness and a cavalier approach to type. But to non–graphic designers, kerning might seem an arcane and unimportant matter. “Looks fine to me,” they might respond when you tell them that you just have to adjust the kerning.
The client may not always be right, but they might well be, in this case.
More problems are caused by overkerning than by lack of kerning. That’s not to say that kerning can’t improve a setting of display type, but overkerning often creates weirder and uglier problems than would be seen by just leaving the type alone—particularly, as is almost always true, when there is kerning present in the font.
Problems with judgement in letter spacing have a long history, even longer than typography itself. The famous inscription at the base of Trajan’s column in Rome, predating Gutenberg by almost 1,400 years, shows examples of this.
In the days of wood and metal, when kerning meant taking a saw or file to your letters, it was something that might or might not be done in large display type. You would not see kerning done in text type, though foundries would sometimes incorporate kerns into their designs; in metal type, a part of a letter might be made to hang outside the type block—and this overhanging part was called a kern—most commonly for italic.
When film typositors, phototype and digital fonts came along in the second half of the 20th century, letterspacing and kerning became completely under the control of the typesetter. While kerning became easy to do, it led to very tightly set type. And this is where overkerning really took hold. There is nothing wrong with very tight overall setting in display type: it is a visual style. The problem arises when people focus on the relationship between individual letters rather than the overall setting.
Since the introduction of digital fonts, most foundries have incorporated kerning values into their typefaces, which is either done visually by the font’s designers, by an algorithmic kerning service, by a font editor’s algorithm or possibly by a combination of the three. The most common programs used for setting type, including word-processing programs, can recognize the embedded kerning. And most of time, that kerning is quite good.
Display type, in headlines and logotypes, is what gets the most attention for kerning—as it should. It’s also where bad kerning interventions are most likely to happen. Most typefaces (but by no means all) have optimized kerning and spacing for smaller sizes, so a bit of kerning will often make a headline better.
But too much attention to kerning can often result in type that is set worse than the default kerns. Another thing that happens occasionally with designer kerning is the tendency to put equal space between letters.
Kerning is, to some degree, subjective. Sure, the letters in a word should appear equally spaced, but designers may not agree on what parts of letter relationships are important to look at for good spacing. The Roots logo is an example. If you focus on the tops of letters in the word, you will gain a different impression of the letter relationships than if you focus on the bottoms.
Kerning in text type
No designer is likely to adjust the kerns in text type, even if they are using a program that allows this, as professional font designers attend carefully to spacing and kerning. However, the font designer may feel pressure to kern text type to look its best in a larger-than-normal setting, because that is the way potential purchasers of fonts often encounter it. This may lead to type that is a bit tight and overkerned. Still, for text, the logical choice is the metric kerns, but even this may not necessarily be the best one because some metrics don’t make sense in every iteration. And when you consider that kerning of text type wasn’t practical until the second half of the 20th century, you might argue that kern pairs in some fonts are hardly necessary.
Optical kerning uses an algorithm that balances the absolute proximity of forms with the white space between them, giving a result usually close to metric kerning. If the kerning in a font is well done, optical kerning won’t be quite as good as it can mess up the rhythm of the type. In the end, it probably won’t make much difference to the reader. Type purists are more likely to insist that the foundry font kerns are the best, even in fonts with less-than-good kerning decisions. The opinions seem to be divided evenly enough on whether to use optical kerning for text to suggest that using either one is acceptable.
Optical kerning is particularly useful when two different fonts are used together (as metric kern pairs only work within a single font) and with fonts that have numerals without proportional figures as an open type option.
Better practices, better results
There are various methods that people suggest for adjusting kerning, such as looking at type upside down, paying more attention to balancing the white space between the positive forms of the glyphs or blurring the type to get a better understanding of the letter relationships. The problem with these methods is that while they may provide some insight into problems, they aren’t the way we actually see type.
In display type, readers see the words as a whole. When reading text, they don’t see the type in any meaningful way; they are focusing on the content. Things that can impair this focus are bad justification and the kind of visual “hot spots” that come from letters being too close together. As often as not, improving kerning can be a matter of reducing kerns rather than increasing them. If you find yourself making big changes after applying metric or optical kerning, you are probably on the wrong track. Perhaps the best way to approach kerning display type is to consider three-character groups. If every three-letter combination is good, then the kerning will be good.
Kerning intervention was once a necessity for good display type. Now, high-quality fonts have kerning that should work in text and needs only minor alterations for display settings. Display type should often be tighter, but reducing the overall tracking is the first action to take and adjustments to kerns after having done so should be minor. All-caps settings might require a closer look as relationships between letters are more likely to be defined by the words that they communicate. Still, instead of holding the attitude that designers must kern type, it is more useful to expect the kerning to already be good and make changes only if they’re minor.
We expect kerning in text type, and usually the foundry kerns will do the job. If they aren’t generally good, it might make sense to choose another typeface, or, failing that, to use optical kerns. Designers shouldn’t feel obliged to kern, as their efforts may result in worse type. You should be able to trust, most of time, font designers’ opinions of how their type should be spaced. What is needed is a critical attitude more than an active one. ca