The monitor is a measuring instrument nobody calibrated

Four ways the screen misleads you

Every editing decision you make is a comparison between what you see and what you want. If the seeing is unreliable, so is every decision, and the results arrive as complaints rather than as a visible fault on your own machine.

Brightness. A laptop at full brightness in a dim room is typically 300 to 500 cd/m². If you edit there, you will brighten the image less than it needs, because it already looks bright to you. The client opens it on a phone outdoors and it is dark. The standard for print work is around 120 cd/m²; for screen work 120 to 160 is sensible. On most laptops that is roughly half the brightness slider, and it will feel dim for a day and then feel normal.

White point and colour cast. Panels ship with wildly different white points, often bluer than the 6500 K standard because blue-white sells in shops. Edit on a blue-tinted screen and you will warm your images to compensate, and every one of them will look orange elsewhere.

Viewing angle. Cheap TN laptop panels shift tone with head position: tilt the lid five degrees and the shadows change. You cannot judge tone on one. IPS panels are stable and most laptops sold now have them; a two-minute test is to open a mid-grey image and move your head.

Ambient light. A window behind you puts reflections on the screen; a warm lamp beside it changes your eye’s adaptation. Your visual system normalises to the room, not the screen. Working in a room with a strong colour on the walls will pull your colour judgement in the opposite direction all day.

What calibration hardware does

A colorimeter (a small puck that sits on the screen, around £110 to £150 for a SpyderX or i1Display) measures what the screen actually emits while software shows it known values. It then produces two things: adjustments to load into the graphics card, and an ICC profile describing the display’s real behaviour so colour-managed applications can compensate.

DisplayCAL, driving ArgyllCMS, is free, open source, and drives almost every colorimeter on the market. The hardware costs money; nothing else has to.

The honest statement about calibrating by eye: you can correct an obvious colour cast with a visual test pattern, and you cannot measure luminance at all without an instrument. Your eye adapts within seconds, which is precisely the faculty that makes it useless as a meter. Visual calibration is better than nothing and is not calibration.

What to check without any hardware at all

Three free tests that catch the worst problems in five minutes.

1. Shadow and highlight steps. Display a greyscale step wedge: a strip of patches at 0, 2, 5, 10 … 95, 98, 100 per cent. You should be able to distinguish 2 per cent from 0 and 98 from 100. If the dark steps merge, your black level or contrast is wrong and you will over-lift shadows in everything you edit.
2. Neutral check. Open a pure 128,128,128 grey filling the screen and look at it against a sheet of white printer paper under daylight. If the screen is obviously bluer or pinker than the paper, you know which direction your judgement is skewed.
3. Uniformity. The same mid-grey at full screen. Cheap panels are often visibly darker in the corners by 10 per cent or more. Judge tone in the middle of the screen, not at the edges.

Check the picture where the picture will live

The most useful habit in this lesson is also the cheapest: look at the finished file on three devices before delivery. Your monitor, a phone, and one other screen: a colleague’s laptop, a TV, a tablet.

Phones deserve particular attention because they do things to your image after you deliver it. Auto-brightness changes exposure judgement continuously. True Tone and Night Shift warm the whole display towards the room. OLED panels push saturation. And a phone is where most of your audience will actually see the work.

If an image looks acceptable on all three, it will survive the fourth. If it looks right on one and wrong on the other two, the one is the outlier.

What calibration does not achieve

It does not make your screen match everyone else’s. Nothing does. The viewer’s phone is uncalibrated, at a brightness they chose, in a room you cannot see, possibly with a blue-light filter on.

Calibrating makes your own decisions repeatable and defensible: the same file looks the same to you next month, two machines in the same studio agree, and when a print comes back wrong the screen is eliminated as a suspect. That is the actual benefit, and it is worth having. Expecting more than that from it leads to the frustrated conclusion that calibration “does not work”.

There is one place it does deliver a hard promise: with a calibrated display and a paper profile, a soft proof predicts a print closely enough to approve it without printing. For anyone doing print work, that alone pays for the puck in a month of wasted proofs.