Face shape from a tape measure

Each of the seven shapes has a page of its own, with a worked set of tape measurements that produces it: the seven face shapes.

How to take the four measurements

A soft tape is easiest. A piece of string and a ruler works just as well, and for the widths it is arguably better, because string does not bend around the face the way a tape wants to. Everything is a straight-line distance, not a distance around anything.

Forehead width

The widest point across your forehead, roughly halfway between your eyebrows and your hairline. Feel for where the forehead stops curving forward and starts turning back toward your temples; that is the point on each side.

Face width at the cheekbones

The widest point across your whole face. Run your fingers back from the outer corner of each eye until you feel the bone ridge under them, then measure between those points. This is the reference every other measurement is divided by, so it is the one worth repeating.

Jaw width

From just below one ear to just below the other, at the corner where the jaw turns up toward the ear. Clench your teeth and you will feel the muscle move; the corner sits just behind it.

Hairline to chin

From where your hair starts at the front, in the middle of your forehead, straight down to the bottom of your chin. Not from the crown. This is the measurement everything below is about.

Why the length is the weak one

The widths share a useful property: both ends of each are places you can feel with your fingers. A cheekbone is a ridge of bone. A jaw corner is an angle you can locate with your thumb. Two people measuring the same face will land within a few millimetres of each other, and a few millimetres on a 140 mm width is under two per cent.

A hairline is none of those things. It is a gradient rather than an edge, it is different on the left from the right, and it moves over the course of a life. Worse, the tool it has to be compared against does not use it at all: the photo version measures down from a reconstructed point on the forehead that sits about 29 mm lower, and that gap varies from person to person by around 8 mm.

Put a 3 mm tape error together with that, and the length ratio comes out with an uncertainty near 0.07. The thresholds separating short from long are 0.10 apart. The error bar covers two thirds of the category.

That is the honest situation, and it is why this page tells you which other shapes your numbers reach rather than printing one and moving on. Every other manual face shape page gives you a single confident answer from the same four numbers. They are not measuring anything you are not.

What survives, and what that means for the answer

The three shapes decided by comparing widths come through cleanly.

Heart is a jaw meaningfully narrower than the forehead. Triangle is the reverse. Diamond is both ends narrow against wide cheekbones. All three are ratios of one width to another, both reliable, and the tool will name them with confidence.

Round, oblong and the oval between them are decided by length over width. Those are the calls the hairline makes shaky, and the tool says so every time rather than only when it happens to notice.

Square needs the two optional vertical measurements, and this is worth explaining rather than hiding. What separates square from round is not a width at all — it is whether the jaw holds its width down the face and turns a corner, or curves away. With only four measurements the calculation reduces to a restatement of the jaw width, which would look like an independent test and be nothing of the kind. The tool declines to run it rather than produce a number that means nothing.

The conversion, in full

Since this page is asking you to trust a number it adjusts before using, here is the adjustment.

Published anthropometric norms put hairline-to-chin over face width at about 1.365 in adult men and 1.357 in women. The photo tool's own typical, measured with its landmarks on real faces, is 1.15. On a face 140 mm wide, that difference is (1.36 − 1.15) × 140, or about 29 mm — which is the distance from the hairline down to the point the mesh treats as the top of the forehead.

Checking it the other way round: an offset between 25 and 30 mm reproduces a hairline ratio of 1.33 to 1.37 across the range of adult face widths, which brackets the published figure. Two routes, one number, so 29 mm is subtracted from your measurement before anything else happens.

Converting the measurement is better than keeping a second set of thresholds for tape users. Two sets would agree on the day they were written and drift apart the first time either was touched — and the drift would be silent, which is the kind of mistake this site tries hardest to design out.

When to use the camera instead

If your hairline has receded, this method does not degrade gracefully, it breaks. The starting point has moved by more than the whole correction. Same for a shaved head, from the other direction.

The camera route also has one advantage that has nothing to do with precision: it measures from landmarks that do not depend on where you decide your hairline is, so two attempts give the same answer. A tape gives you whatever you did that time. If you want a number you can come back to, that matters.

Nothing is uploaded either way. The photo tool runs the same arithmetic in your browser and the image never leaves your device — the reason to use this page is that you do not want to use a camera, not that the camera version sends anything anywhere.

Face shape categories are a convenience rather than a fact — faces vary continuously and the boundaries are lines drawn through a continuum. This tool reports the measurements, the thresholds and the uncertainty so you can judge the classification instead of taking it on trust. It does not score, rank or assess anyone's appearance.

Questions people ask

How do I find my face shape without a photo?

Four measurements with a tape or a piece of string and a ruler: the widest point across your forehead, the widest point across your face at the cheekbones, the width of your jaw from below one ear to below the other, and the distance from your hairline straight down to the bottom of your chin. Enter them above in millimetres. Two more vertical measurements are optional and are what separate a square face from a round one.

Is measuring with a tape as accurate as using a photo?

For the widths, near enough — both ends of each are places you can feel with your fingers, and a few millimetres of error barely moves the ratio. For the length, no, and the reason is specific: the photo tool measures from a fixed point on the forehead, while a tape has to start at the hairline, which sits about three centimetres higher and varies between people. That uncertainty alone puts roughly ±0.07 on the length ratio, against category thresholds only 0.10 apart. So the shapes decided by width come out reliably and the ones decided by length do not.

Which face shapes can this tell apart, and which can it not?

Heart, triangle and diamond are decided by how the forehead and jaw compare with the cheekbones — all widths, all measurable with a tape, all reliable. Round, oblong and oval are decided by length over width, which is the measurement the hairline makes unreliable, so the tool names the other shapes your numbers could also produce instead of picking one and sounding certain. Square and round need the two optional vertical measurements; without those, the calculation would only be restating the jaw width.

Where exactly is my hairline for this?

Where the hair starts at the front, in the middle of your forehead, not at the crown. Measuring to the top of the head is the commonest mistake and adds five centimetres or more, which is enough to turn an ordinary face oblong — the tool checks for it and says so rather than producing a confident wrong answer. If your hairline has receded, this method does not work, because the starting point has moved by more than the calculation allows for.

Why does the length get converted before it is used?

Because the thresholds were calibrated on what the photo tool measures, and the two instruments measure different things on the same face. The mesh's top-of-forehead point sits about three centimetres below the hairline, so a tape reading has to have that subtracted before it can be compared. The offset was worked out two ways that agree: from published anthropometric norms, and from what offset reproduces those norms across the range of adult face widths. Reusing one set of thresholds beats maintaining two that would drift apart.

What if I get a different answer than the photo tool gave me?

Trust the photo. It measures from fixed anatomical landmarks rather than from a hairline, it averages across the frame, and it corrects for head turn. If the two disagree, the difference is almost always in the length, and almost always because the hairline is not where the calculation assumed. The overlap is worth looking at rather than ignoring: if a tape and a camera put you in different categories, you are close to a boundary, which is more useful to know than either answer on its own.

Do I need someone to help me measure?

It is easier with help, and more accurate, particularly for the jaw width and the hairline-to-chin length where you cannot see what the tape is doing. Measuring in a mirror works but tends to tilt the tape. If you are alone, take each measurement three times and use the middle value — repeated readings expose a slipping tape far better than care on a single one does.