Why font hinting still matters in 2026
Retina screens were supposed to make font hinting obsolete. They didn't. The reasons hinting still matters tell you something about how typefaces survive contact with real screens — and why some of them don't.
Retina screens were supposed to make font hinting obsolete. They didn't. The reasons hinting still matters tell you something about how typefaces survive contact with real screens — and why some of them don't.
For most of the last fifteen years, designers have been told that font hinting is a dying craft. The argument went: high-density screens (Apple's Retina, Samsung's AMOLED, every modern laptop) have so many pixels that the typeface no longer has to be coaxed onto the grid. The hinting instructions — the tiny rules a type designer writes to tell the rendering engine which pixels to light up at small sizes — were yesterday's problem.
This is mostly wrong. It was wrong in 2014 when people first started saying it, and it is wrong in 2026 in a more specific way that's worth understanding. Hinting still matters. The reasons it matters are different now, and the typefaces that ignore hinting have a particular quality that you can spot once you know what to look for.
A typeface, in its raw form, is a collection of vector outlines — mathematical curves that describe the shape of every character. When you display a typeface at a particular size on a screen, the rendering engine has to convert those curves into pixels. At 96 pixels per em (a fairly typical body-text size on a non-Retina display), you have very few pixels to play with. A capital E might be 12 pixels tall. The horizontal arms of the E are each one or two pixels.
If the rendering engine just rounds the curves to the nearest pixel, you get a mess. The arms of the E might land between pixels and get smeared across two rows, going from one solid pixel-row into two half-strength pixel-rows that look fuzzy. The verticals might be one pixel wide on the left and two pixels wide on the right, depending on where the outline happens to fall on the grid. The whole letter looks lopsided and slightly ill.
Hinting is the set of instructions the type designer attaches to the typeface to tell the rendering engine how to handle these cases. This stem should always be exactly two pixels wide. These two strokes should be the same weight as each other. This serif should snap to the pixel grid even if the math says it should land between pixels. Good hinting is invisible — it just makes the typeface look correct. Bad hinting is visible everywhere.
The argument that Retina killed hinting goes like this: at high pixel density, you have so many pixels per em that the rounding errors don't matter. A capital E might be 30 pixels tall instead of 12. A one-pixel rounding error becomes invisible because there are so many other pixels around it.
This is true at large display sizes. It is much less true at body text sizes. A 16px body paragraph on a Retina laptop display still has cap-heights in the 11-12 actual-pixel range (because the device pixel ratio doesn't actually double the type's grid resolution for rendering quality the way it doubles physical resolution). The same rounding problems apply, just at a smaller absolute size.
And — crucially — most of the world is still not on Retina. Look at any analytics dashboard. Cheap Android phones, kiosk displays, in-car screens, point-of-sale terminals, library computers, the dashboard at the gym, the screen at the airport, the projector at the office. Every one of them is running typefaces at low device-pixel-ratio at small sizes. The browsers that render them care about hinting. The users reading from them care about hinting whether they know it or not.
You can spot a well-hinted typeface against a poorly-hinted one without any technical training, once you know where to look.
Stem consistency. Set the typeface in lowercase at 14px on a low-density screen. Look at the verticals of the m, n, h, b, d, p, q. They should all be the same width as each other. In a poorly-hinted typeface, some of them will be a pixel wider than others. The eye reads this as the typeface being slightly drunk.
x-height stability. The lowercase letters should all sit on the same baseline and rise to the same x-height. In a poorly-hinted typeface, the rounded letters (o, e, c, a) will sit a half-pixel higher or lower than the flat-bottomed ones (n, m, h). This is called overshoot — designers add overshoot deliberately at large sizes so curved letters don't look smaller than flat ones — but at small sizes, hinting should pull the overshoot back so the line looks even.
Color (in the typographic sense). A paragraph of well-hinted type at body size should look like an even gray block from across the room. A poorly-hinted typeface will have darker words and lighter words depending on which letters happen to round well at that size. The unevenness reads as visual noise.
The reason a lot of contemporary typefaces ship with token hinting (or auto-hinted defaults from the font tooling) rather than carefully hand-tuned hinting is simple: hinting is expensive. A skilled hinter can take weeks per weight to do it right. A small foundry releasing a six-weight family with full italics and a complete language extension is looking at months of hinting work after the typeface is otherwise done.
For a foundry whose customers are mostly buying the typeface for headlines and brand work — where the typeface will mostly be set huge — the hinting investment is hard to justify. It pays off in a context the foundry's customer doesn't usually care about.
This is why a lot of beautiful display typefaces released in the last five years look slightly off when you actually try to use them as body text. They were drawn for the showings, not for the small sizes. Hinting is the gap.
If your project is going to use a typeface at body text sizes — anywhere on the web, in app interfaces, in documents people will actually read — hinting is a real consideration. It is not a craft of the past. It is the difference between a paragraph that's pleasant to read and one that quietly fatigues the reader without them being able to say why.
Some practical advice:
Hinting will eventually matter less, in the same way the difference between a CRT and an LCD eventually mattered less. We will, eventually, be on screens dense enough that the rendering engine can ignore the grid. We are not there yet. The typeface you spec today will be read on screens that exist today.
The craft of hinting is one of those quiet, invisible disciplines that only ever announces itself when it's missing. The typefaces that hold up at every size, on every device, in every reader's hand — those typefaces have hinting in them. Notice which ones do, and the world of typeface choice gets smaller, more honest, and a great deal more useful.
Be the first to mark this page