Why always-on displays dim, and what burn-in actually is

Look at a phone lock screen at night and it is doing four distinct things at once, not one. It has dropped the brightness, thinned the type, discarded most of the interface, and — if you watch it for an hour — quietly moved the clock a few pixels. Only the first of those is about your eyes. The other three are about the physics of OLED panels, and they are the reason your phone's screen still looks new after three years of doing this every night.

An OLED pixel is a tiny lamp

An LCD has one backlight behind the whole panel and a layer of liquid crystal that blocks it selectively. Black on an LCD is the backlight still blazing away behind a shutter that is doing its best. This is why blacks on cheap LCDs look grey, and why an all-black LCD image draws almost exactly as much power as an all-white one.

OLED has no backlight. Every subpixel is its own light source. Black is not a colour being displayed — it is a lamp switched off, drawing no current at all.

Two consequences follow, and everything else in always-on display design falls out of them. A dark image on OLED is genuinely cheaper to show than a bright one. And a lamp that is on is a lamp that is wearing out.

Burn-in is uneven ageing

Organic LED emitters dim as they are used. The decline is gradual and roughly proportional to cumulative light output — hours multiplied by brightness. A panel where every pixel had shown the same average brightness for its whole life would dim uniformly and you would never notice.

Screens do not work that way. A status bar sits in the same place for years. A navigation bar, a keyboard, a game's heads-up display — all static, all in fixed positions. Those pixels accumulate far more lit hours than the ones around them, so they fade further. Show a flat grey afterwards and the ghost of the status bar appears, not because anything is stuck on, but because that region can no longer get as bright as its neighbours.

That is burn-in: not an image burned into the glass, but a permanent difference in how much light different regions can still produce.

It is also uneven by colour. Blue OLED emitters are the least efficient and degrade fastest, so a heavily-aged region does not simply darken, it shifts warm. Ghosts often read as faintly yellow.

Why thinning the type beats fading it

Here is the part that is easy to get wrong. If you want an always-on clock to be dimmer, the obvious move is to lower its opacity. Fading white text to 40% grey certainly looks dimmer.

But consider what the panel is doing. Every pixel that made up the numerals is still lit — just less brightly. The same area of screen is still ageing, only slower.

Now consider making the strokes thinner instead. A variable font at weight 145 instead of 340 lights meaningfully fewer subpixels. The pixels that fall away are not dimmed, they are off — drawing nothing and ageing not at all. The clock reads as lighter and quieter, and the wear is concentrated on a smaller area.

This is why phone lock screens use a thin weight for the always-on state and a heavier one when you wake the device. It looks like a stylistic choice. It is a power and longevity decision that happens to look good.

The strongest version does both: fewer lit pixels, each less bright.

Shedding everything that is not the time

The same logic explains why always-on modes discard so much. Notification icons, battery indicators, seconds counters, wallpaper — each is more lit area for information you did not ask for while the phone sits face-up on a desk.

An always-on display is not the normal interface at lower opacity. It is a deliberately smaller set of things, chosen because each one costs lit pixels. When a clock drops its seconds indicator and its controls on the way into always-on mode, that is the reasoning.

The pixels move

The last mitigation is the one nobody notices, which is the point. Phones shift the position of an always-on clock by a few pixels on a slow cycle — typically a minute or so between steps, following a small loop.

The purpose is to stop any one pixel from sitting under the same glyph edge for hundreds of hours. Spread across a slightly larger area, the wear per pixel drops and no sharp boundary forms. A ghost needs a hard edge held in one place; keep the edge moving and it never establishes.

The movement has to be slow and small enough to be invisible. Thirteen pixels over an hour is imperceptible unless you are looking for it, but it is enough to blur the boundary across months.

What this means for a clock on your own screen

If you leave a clock running on an OLED panel — a phone as a nightstand clock, an OLED laptop on a desk — the four measures apply to you exactly as they do to a phone:

• Use a dark theme, and one that goes to near-black rather than dark grey. The difference between #111 and #000 is the difference between dim lamps and no lamps.
• Turn the brightness down. Ageing scales with it directly.
• Prefer thin type over faded type.
• Let the image move, if whatever you are running supports it.

If your screen is an LCD, none of this applies in the same way. LCDs show temporary image retention, where a faint ghost persists for minutes or hours and then clears. It is annoying, not permanent, and it has a different cause — trapped charge in the liquid crystal layer, not worn-out emitters. An all-black LCD also saves you nothing on power, because the backlight is still on.

Knowing which panel you have decides how much of this you need to care about. Phones from about 2017 onward and iPhones from the X onward are mostly OLED. Desktop monitors mostly are not, yet.

The clock does all four of these in its always-on mode. Or read about setting up a spare phone as a bedside clock.