Your season, with the lighting taken out of it

The problem every colour tool has, and most ignore

Ask someone to photograph their face and analyse the skin pixels, and what you actually measure is the skin multiplied by the light falling on it. A tungsten bulb adds orange. An overcast window adds blue. A bathroom LED can add green.

The effect is not subtle. The colour shift from a warm indoor bulb is comfortably larger than the difference between a Cool Summer and a Warm Autumn. This is the real reason people run three colour-analysis apps and get three answers: the apps are measuring the rooms.

Using the eye as a grey card

Photographers solve this with a grey card — a patch of known neutral placed in the scene, which tells the software what white looked like under that light. You almost certainly do not own one.

You do own two, roughly, in the whites of your eyes. The sclera is close to neutral, it sits in exactly the same light as the rest of your face, and it appears in every usable portrait. Scaling the red and blue channels until the sclera reads neutral removes most of the cast before any skin measurement happens.

It is the same move as the PD tool using the iris as a millimetre ruler, and it has the same appeal: no props, nothing to buy, and the reference is guaranteed to be in the same conditions as the subject.

The honest limits: sclera is faintly blue in youth and yellows somewhat with age, it sits partly in eyelid shadow, and a strongly coloured light cannot be fully undone by any correction. So the tool samples only the brightest sclera patches, reports the size of the correction it applied, warns above 18 percent, and refuses outright above 40 rather than hand you a confident season derived from a lamp.

The four things actually measured

Skin hue, as an angle in the CIE Lab colour space. All human skin is some version of orange, sitting in a band from roughly 45 to 75 degrees, so the question is never whether your skin is orange but where inside that band it falls. Below about 54 degrees it leans pink and reads cool; above 62 it leans golden and reads warm; between the two is genuinely neutral, which is more common than the internet suggests.

Skin lightness, the L value, from 0 for black to 100 for white. This is the depth axis, and it is independent of hue: deep skin can be cool and light skin can be warm.

Eye chroma, how saturated the iris is. Bright, saturated eyes push toward the clear palettes, which is the observation the whole clear-versus-soft axis rests on.

Contrast, the lightness gap between hair and skin. Dark hair on light skin is high contrast and carries strong colours; mid-brown hair on medium skin is low contrast and gets overwhelmed by them.

Why hair is a question rather than a measurement

Because sampled hair colour is so often not the person's hair colour. It might be dyed, tied back, cropped out, greying unevenly, or lit from behind. Measuring it and presenting the number as data would be dressing up a guess.

So we ask, and we use the answer for one thing only: the contrast calculation. If you dye your hair, answer with the colour you actually wear, since that is what the palette has to sit against.

The twelve seasons

Four families, each split three ways by whichever characteristic dominates — temperature, value or chroma.

How much weight to put on the answer

The measurements are real. Hue, lightness, chroma and contrast are well-defined quantities in a standard colour space, computed the same way every time, and this tool shows all four so you can check whether your result was a clear case or a near miss on a threshold.

The seasonal system sitting on top is a framework from the styling industry, developed in the twentieth century and refined by practitioners rather than derived from research. It is a useful way to group colours that tend to sit well against a particular colouring. It is not a fact about you, and where it disagrees with a colour you know suits you, believe yourself.

One practical note: a professional in-person analysis costs anywhere from £100 to £500 and involves draping real fabric against your face in controlled light, which remains more reliable than any photograph. What this gives you is the measurement part, for free, with its error sources stated.

This tool analyses a photograph on your device. No image, sample or result is transmitted or stored. Results depend heavily on photo quality, and the seasonal framework is a styling convention rather than a scientific classification.

Questions people ask

Why does this need the whites of my eyes?

As a colour reference. A photograph carries whatever colour cast the room had, and a warm bulb will turn a cool complexion warm. The sclera is approximately neutral and sits in exactly the same light as your face, so correcting the image until the sclera reads neutral removes most of the cast before any skin is measured. It is the same idea as using the iris for scale in the PD tool: a reference you already have.

Why do other colour analysis tools give me a different season each time?

Almost always because they ignore white balance. Photograph the same person under a tungsten lamp and again by a window and the raw skin pixels differ enormously, so the answer flips. Any tool that returns a season without telling you how much colour cast it corrected is reading the room's lighting as much as your face.

Why do I choose my hair depth instead of the tool measuring it?

Because measured hair colour is frequently not your hair colour. It may be dyed, tied back, out of frame, greying unevenly, or simply in shadow. Rather than sample something unreliable and present it as data, we ask, and we use your answer only for the contrast calculation.

Is seasonal colour analysis scientific?

The measurements are: skin hue, lightness, eye saturation and contrast are real quantities in a real colour space, and this tool shows you all four. The twelve-season system built on top of them is a framework from the styling industry, not a finding from research. It is a useful organising idea for which colours tend to sit well together against a given colouring, and it should be held about that loosely.

Can I trust a result from a phone selfie?

With caveats. Use daylight near a window, no coloured lamps, no filters or beauty smoothing, and no makeup on the cheeks. The tool reports the size of the cast it corrected — under about 18 percent is a clean photo, and above 40 percent it refuses rather than guess.

What about makeup and tan?

Both move the answer. Foundation in particular is usually chosen to a shade rather than matched to the skin, so it can shift the hue by more than the difference between two adjacent seasons. A fake tan does the same. Analyse bare skin, and if you are currently much more tanned than usual, expect the result to read warmer than your baseline.