Undertone, measured rather than guessed at

Surface tone and undertone are different things

Surface tone is how light or deep your skin is, and it moves — with sun, with season, with a flush. Undertone is the colour sitting beneath it, and it barely moves at all.

That is why depth and undertone have to be measured separately. Very fair skin can be warm, very deep skin can be cool, and olive appears at every depth there is. Treating them as one axis is why so many people end up with foundation that matches their lightness and fights their colour.

Why the usual tests fail

Almost everyone has been told to look at their wrist veins. Blue means cool, green means warm. It is memorable, and it does not measure what it claims to.

What you see through skin is a scattering effect — shorter wavelengths scatter back more, which is why veins look bluer than the blood in them actually is — and that effect is filtered by the thickness and pigmentation of the skin above. Two people with identical undertones and different skin thickness will report different colours. On deeper skin tones the veins often cannot be seen clearly at all, which makes the test not merely unreliable but unusable for a large fraction of people.

The gold-versus-silver test is better, because at least it asks about the thing in question. But it asks you to judge it under whatever light you happen to be standing in, and lighting is precisely the variable that decides the answer. Most people also answer from habit — the metal they have worn for twenty years looks right because it looks familiar.

What this measures instead

Skin is sampled from several points across the cheeks and forehead, the median taken so a blemish or a shadow cannot drag the result, and the values converted into CIE Lab — the colour space built so that distances in it correspond to perceived differences.

Two numbers then do the work. The hue angle says where your skin sits in the narrow orange band all complexions occupy: below about 54 degrees it leans pink, above 62 it leans golden, and in between is genuinely neutral. The ratio of yellowness to redness catches the case the hue angle alone misses.

Olive, and why it is flagged rather than declared

Olive skin carries a green cast that damps the redness while leaving the yellow in place. The practical consequence is one of the most common shade-matching failures there is: someone with olive skin is told they are warm, buys a golden foundation, and it turns orange within hours because the product supplies warmth the skin already has while failing to supply the green that damps it.

So it would be useful to detect. The difficulty is that hue angle carries no information about it. The first version of this tool separated olive using the ratio of yellowness to redness, which turned out to be exactly the tangent of the hue angle — the rule was "hue above 69 degrees" in disguise, and it relabelled every strongly warm complexion as olive.

The only axis independent of hue is chroma, how saturated the colour is, and olive does tend to measure less saturated at a warm hue. But that relationship is weak enough, and varies enough with depth and with the camera, that declaring olive from one photograph would claim more than the measurement supports.

So the tool reports three undertones and raises olive as a prompt when your reading is warm-leaning but desaturated, along with the test that actually settles it: whether a golden foundation turns orange on you within a few hours. That is the honest shape of what a photograph can tell you here.

The lighting problem, and the eye trick

A photograph records your skin multiplied by the light falling on it. A warm bulb shifts a complexion further than the entire gap between cool and warm, so any tool that samples raw pixels is largely measuring the room.

The fix is a neutral reference in the same light as the subject. Photographers use a grey card; you have the whites of your eyes, which are close enough to neutral and guaranteed to be lit identically to your face. Scaling the channels until the sclera reads neutral removes most of the cast before any skin is measured.

It has limits. Sclera is faintly blue in youth and yellows with age, it sits partly in eyelid shadow, and no correction fully undoes a strongly coloured light. So the tool reports the size of the correction it applied, warns above 18 percent, and refuses above 40 rather than hand you a confident answer derived from a lamp.

Using the answer

For foundation, undertone picks the letter and depth picks the number. Cool ranges are usually marked C or P, warm W or G, neutral N, and olive is increasingly given its own line. Match depth along the jaw in daylight, then check the letter by seeing whether the shade turns orange or grey after a few hours — orange means too warm, grey means too cool. That drift is a better test than anything you can see at the moment of application.

For everything else the effect is real but mild. Metals, whites and the colours listed above are conventions that tend to sit well against a given undertone, not rules, and a colour you know suits you beats any of them.

This tool analyses a photograph on your device and transmits nothing. Results depend heavily on lighting, and the guidance offered reflects conventional colour practice rather than measurement.

Questions people ask

Does the vein test actually work?

Not reliably. The idea is that blue veins mean cool and green veins mean warm, but what you see is a blue-shifted scattering effect through skin, filtered by the thickness and pigmentation of the skin above it. Two people with the same undertone and different skin thickness will see different colours, and deeper skin tones often cannot see veins clearly at all. It is a memorable rule that happens not to measure what it claims to.

What about the gold versus silver test?

It is closer to useful, because it asks you to judge the same thing the classification is about — whether a golden or a bluish cast sits better against your skin. The problem is that it is a judgement made under whatever lighting you happen to be in, which is the exact variable that ruins every visual method, and a strong preference for one metal usually reflects habit rather than complexion.

What is an olive undertone, and why won't the tool just tell me?

Olive is a green cast that damps the redness in the skin without removing the yellow, and it occurs at every depth from very fair to very deep. We flag it as a possibility rather than declaring it, because hue angle carries no information about olive at all — the only independent signal is chroma, and the relationship is too weak to call from a single photograph. The test that settles it is behavioural: if a golden foundation turns orange on you within a few hours, work from olive and neutral ranges rather than warm ones.

Can undertone change?

Surface tone changes with sun exposure, season, flushing, medication and age. Undertone is far more stable, but a strong tan shifts the measurement warmer because it is genuinely adding pigment above the same underlying colour. Measure bare, unflushed, untanned skin for a baseline, and re-measure after a summer if you want the shade you need right now.

Why does this need a daylight photo?

Because a photograph records your skin multiplied by the light falling on it, and a warm indoor bulb shifts a complexion by more than the entire gap between cool and warm. The tool corrects using the whites of your eyes as a neutral reference, reports how large a correction it applied, and refuses outright above 40 percent rather than guess.

Does makeup affect it?

Considerably. Foundation is chosen to a shade rather than matched to skin, so it can shift the hue further than the difference between two adjacent categories. Measure bare-faced. Even tinted moisturiser and SPF with a cast will move the result.