Glasses for Diamond Face Shape: Five Frame Numbers, Three Outlines and One Subtraction

Five rows, three outlines, one working width

• Four of the five rows are flat across three columns, and they are flat for a mechanical reason: those four are set by how wide a skull is behind the ears, how wide a nose sits between the eyes, and how far apart the pupils are. A face-shape result contains none of those three measurements. Every size chart checked for this page hands the same front to a diamond face and a heart face of the same width.
• The row that does move separates two of these three outlines by 1 mm (0.04 in), which is not a difference anybody can shop against. This site's diamond centre on that row is interpolated from three photographs; four of the seven outlines carry a measured sample of five faces or more, and the remaining three sit on less than that — diamond at three photographs, heart at one, triangle at none — so their centres come from the published ordering rather than from measurements. A millimetre of corridor is well inside what a sample of three can hold. The correct reading of the row is that diamond and heart share a vertical budget.
• The interesting column is the one not printed. A round face at the same 136 mm cheekbone line stops at 44 mm (1.73 in), four millimetres below the diamond row, and 4 mm is most of the distance between the heights the trade actually cuts: the charts that give lens heights at all print them at 40, 43, 46 and 50 mm.
• The depth row starts at 30 mm (1.18 in) rather than at the 40 mm (1.57 in) where the fashion bands open, because spectacle fronts are cut shallower than sun fronts. That single gap in the floor is why this page is not the sunglasses page in the same family: the row a lens with power in it has to answer to is not the row a tint answers to.

What the outline actually buys

Take the disappointment in that table straight, because it is the useful part. Being diamond changes one of five numbers, changes it toward more room rather than less, and then the change is smaller than the uncertainty in the label itself. What a diamond result entitles you to at a shop counter is not a wider frame — you are already widest at the cheekbone, and width there is decided by the tape, not the taxonomy. What it entitles you to is a deeper lens than a round face can wear, and that is a permission you then have to clear with somebody else's arithmetic.

That is a different finding from the oval page's, where the shape word turns out to move nothing and the practical object collapses into a size ladder picked off one face-width measurement. Here the shape word moves something real — this site's corridor ceilings put a diamond face 4 mm (0.16 in) deeper than a round one at the same width — and the thing it moves is exactly the thing a strong prescription refuses to pay for. Section four works that out in millimetres.

And if what actually brought you here was a result you did not trust, the four flat rows are the reassurance worth having: a disagreement between diamond, oval and heart changes nothing about what you order. Width, bridge, temple and lens width come off your measurements whichever of the three the classifier settles on. The label is worth having for what it says about hair, contour and chin. It is not worth re-measuring three times over a frame.

Two numbers on the arm, two on you

Everything in this article is arithmetic on four inputs, and only two of them are printed on the frame. The three figures inside the temple arm name a size rather than a wearer: lens width, bridge width, temple length, in that order, with either a box symbol or a dash between the first two, so 52□18 and 52-18-140 are the same front. Rows two, four and five of Table 1 are the published ranges those stamps fall inside — ranges written for adults in general, and, as far as this page could find, for no face shape in particular.

The two that decide are yours. The width across the tops of the cheekbones you can take in fifteen seconds with a soft tape and your hair pushed back. The distance between your pupils comes out of a refraction and out of nothing else; two published adult readings disagree about where the average sits — a hospital health article published April 2026 gives 50 – 70 mm (1.97 – 2.76 in) with an average of 63 mm (2.48 in), while an eye-health reference updated August 2025 puts adults at 60 – 65 mm (2.36 – 2.56 in) — and both are honest about the populations they measured.

Now add two of the stamped numbers together. Lens width plus bridge width is the distance between the two optical centres of that front as manufactured, which is the frame's pupillary distance: 46□16 gives 62 mm (2.44 in), and 52□18 gives 70 mm (2.76 in). Measured across the whole front, that second frame spans 122 mm (4.80 in) of lens and bridge, and roughly 128 mm (5.04 in) once both end pieces are counted. The 128 mm is the number a shop tape reads; the 70 mm is the number your prescription has to satisfy, and they move together only through the lens width.

Subtract your own pupillary distance from the frame's and halve the result. At 52□18 on a 62 mm pupillary distance: (70 − 62) ÷ 2 = 4 mm (0.16 in) per eye. A laboratory-side technical guide last updated May 2026 publishes that figure's ceiling at 5 mm (0.20 in) per eye. This page works to 3 mm (0.12 in) instead, which is tighter than anything published and is stated here as this site's own target, not as a standard. A negative answer means the two centres have to travel outward rather than inward, and the blank formula in the next section was built for the inward case.

Notice what is absent from that subtraction. Your face is in it twice — once as the cheekbone line that sets the front's total width, once as the two pupils that set the frame's — and your outline is in it not at all.

The blank, which is where the depth goes

A finished single-vision lens is cut out of a round semi-finished disc, and the disc that constrains every calculation below is the largest one commonly stocked, at 75 mm (2.95 in). The condition the disc imposes is a minimum-blank one, and this site has already opened that formula up and worked it across seven real frames on the eyeglasses page for a round face, so it is quoted rather than re-taught here; only its output is used below.

Put the three published figures against each other at this page's working inputs. A 62 mm pupillary distance under a 5 mm per-eye ceiling allows a frame distance of 72 mm, which spends 10 mm of the disc; 75 − 10 − 2 leaves 63 mm (2.48 in) for the diagonal the boxed lens has to sit inside, the quantity the round-face page defines and works through. That 63 is the entire budget, and it prices width and depth against one another at a fixed exchange rate: because the diagonal is the hypotenuse, the first millimetres of depth cost a little width, and the last ones cost a great deal.

• Read the last column, not the second. The first five rows are comfortable: a lens the disc can give up is also wide enough to carry the front past the cheekbone line. The sixth is not. At 48 mm of depth the widest lens the blank allows is 40.8 mm, which is below the 42 – 45 mm of the smallest rung on the published adult size ladder — so the depth this site publishes for your outline is reachable, and the width this face needs is not, at the same time. One of the inputs above has to give before a purchase does.
• The crossover sits at 45 mm (1.77 in) of depth, which is not a rung the trade cuts. Between the 43 mm and 46 mm heights the front stops covering the face it was sized for, so the deepest stocked rung that satisfies both the disc and the cheekbone at these inputs is 43 mm (1.69 in) — a lens height you can order, and the number to shop against. It is one rung above the 40 mm (1.57 in) depth ceiling the round-face styling pages publish for their own outline, and one rung below the 46 mm that this site's corridor says a diamond face could wear if nothing else were in the way.
• The material of the front moves that crossover further than the outline does, because the end pieces are the only width in the sum that is free. The allowance needed to reach 136 mm once the blank has capped the lens comes out at 11 mm (0.43 in) at 34 mm of depth, 14 mm (0.55 in) at 38, 16 mm (0.63 in) at 40, 19 mm (0.75 in) at 44 and 24 mm (0.94 in) at 48. Against the medians published in August 2026 from 3,942 retailer frame records — acetate 20 mm, plastic 16, titanium 15, metal 13 — the deepest depth each material's end pieces can carry is 44 mm for acetate, 40 for plastic, 38 for titanium and 34 mm (1.34 in) for metal. Set those four against the 43 mm rung in the bullet above and they are not the same kind of number: 44 mm is what a material supports, while 43 mm is the deepest height anybody stocks, so on the shelf an acetate front and the blank agree and stop together. Read down the four heights the trade prints — 40, 43, 46, 50 — and only acetate reaches 43, plastic makes its ceiling at 40, and titanium and metal finish below the shallowest rung in that set. Those are medians for a material, and an individual frame sits on either side of its own.
• Your pupillary distance moves it in the direction people do not expect: the wider your pupils, the more frame distance you may spend before the decentration ceiling stops you, and so the more width survives for the front. Swap the working input from 62 mm to 66 mm (2.60 in) and an acetate front reaches the full 48 mm; swap it to 58 mm (2.28 in) and it reaches 40 mm. Eight millimetres of pupil reading is eight millimetres of usable lens height.

What a deep lens does at the bottom of a rim

The corridor this site measures runs from the brow to the top of the cheekbone. On a diamond outline the ceiling of it is high and its floor sits at the widest line on the face, which is a combination no other row in the table has: the lens that clears the corridor on paper arrives at the cheek from directly above the bump that defines the shape. So a depth figure can pass Table 2 and still fail on the person. The failure has an order to it. At rest the rim floats. Then you talk, or eat, or grin at something, and the mid-face lifts under the lower edge; the millimetre of clearance is gone, the rim rides the cheek, and the front levers up off the nose in answer. That lever lands as pressure between the eyes, which is why rim height gets misdiagnosed as a nose-pad problem: the pads are the part you can feel.

Two geometric properties decide whether a given front does this, and neither appears on a stamp. The first is how far the bottom of the lens leans toward the cheek when the frame is set level — the steeper that angle, the closer the lower rim runs to the face at precisely the height where this outline is broadest. The second is whether the bottom edge of the shape is cut away at all: fronts exist with a flattened or scooped lower rim, and that scoop is clearance rather than decoration. Deep lens, steep lean, no scoop is the combination that produces the complaint, and taking any one of the three out usually makes the same frame wearable. That is this page's reading rather than a published rule, and it is testable in a shop in under a minute: put the frame on, grin as hard as you can manage, and have somebody watch the lower edge from the side, because from your own angle the rim is the last thing you look at.

The other thing depth does is lengthen the part of a minus lens that is thickest. A nearsighted lens is thin at its centre and thick at its perimeter, so the heavy edge follows the outline you picked, and a deeper shape gives that edge more bottom — which is where a face this wide is also carrying the most of the frame's mass. Published guides begin describing a wide, deep front as a technical problem rather than a matter of taste somewhere past −6.00 D, and the same forces are running quietly by −4.00 D. No English source this page could find attaches a millimetre of permitted depth to a dioptre figure, let alone to a face shape, so the instruction is to carry your own number into Table 2 and read its last row before concluding that the corridor is your limit.

Which is the part of this that inverts the usual shape advice. If your prescription is strong and your face is a diamond, the row your outline opens is the row your prescription shuts. The four millimetres of extra corridor this shape holds over a round one are not a benefit you receive; they are a permission you hold while the blank decides whether you may use it.

What a second pair teaches

A first pair is bought blind. A second is bought from what the first one did, and six lessons follow so consistently from that that they are worth writing down before you place an order rather than after.

Look at the lenses of a pair that came out wrong, from the front, under a lamp. If the inner edges — the sides toward your nose — are thin and the outer ones heavy, both centres were travelled a long way inward to find your pupils, and the front's pupillary distance sat well above yours. You can make that diagnosis with no instrument, on glasses you already own, and it is the reason the subtraction in section three is worth doing before an order rather than after a delivery. It reads like a workmanship fault. It is not. It is what a narrow pupil reading in a broad front produces when the job is even done correctly.

Concentric rings around the lens edge, brightest when you catch yourself in a mirror looking downward, are the same fact seen from the other side: a thick edge on a deep shape, refracting light out of its own perimeter. The previous pair did not do it because the previous pair was narrower, or shallower, or both.

The highest index in the case is the weakest lever on the table. Buying index to rescue a frame that is too large for your pupils trades a thickness problem for an optical one, because dispersion rises as index climbs, and the rainbow that appears at the edge of a 1.74 lens is not something anyone will agree to remake for free. Frame size and decentration move edge thickness by more than an index step does, and they cost nothing.

Frames photographed with the demonstration lenses left in are a different object from frames photographed with them out. A front shot without glass discloses nothing about depth, reads narrower across than it will wear, and is the commonest way a specification gets filled in from a picture. Ask for the depth in millimetres. If it is not on the listing, that silence is information about the frame.

Remaking is not repairing. Moving an existing pair of lenses into a smaller front, or a fresh pair into a frame that changes the centring, relocates both optical centres and can drag the vertical fitting height and a cylinder axis after it — and on an astigmatic prescription a shifted axis is not a cosmetic matter. A re-grind is a second pair built on somebody's second set of tolerances. Decide accordingly.

And the vertical centre is missing from every table here. Fitting height — how far up the lens your pupil sits — is what a deep front and a progressive depend on, it is the measurement the busy counter skips, and this site could not find an English millimetre for it tied to any frame size, so none is printed. If the lens you want is deeper than the rungs in Table 2 clear, or if anything in it is progressive, ask for that reading by name rather than hoping the shape absorbs it.

This page, and the three neighbours that are not it

Four pages in this family answer four questions, and reading them in the wrong order is how an afternoon goes.

This one is the diamond outline with power in the lens: which of the five rows your face moves, and how much of that movement a lens blank will fund.

Glasses for a round face is the same five rows on an outline where four constraints pull against one another — width reaching the cheekbone, depth staying off it, decentration under its published maximum, and a blank that may refuse the combination outright. It is where the decentration and minimum-blank formulas are worked across seven real frames. Read that one if you want the arithmetic shown rather than asserted.

Sunglasses for a round face is lens height as a corridor, published across five outlines including this one, for eyewear with no power in the lens. Its second table is where the 48 mm figure used throughout this page is computed from the rule set and set against the heights the trade stocks. Read it first if you are buying sun frames rather than spectacles.

Glasses for an oval face draws the opposite lesson from Table 1 — that on an oval outline the shape word moves almost nothing, and the useful object is a size ladder picked off a face-width measurement. Read it if your result sat between oval and diamond and you want to know which of the two labels is worth acting on.

The diamond outline guide is where the five readings that produce the label are defined. It prints no frame millimetres at all, on purpose.

Before you order: three measurements and two subtractions

• Take the cheekbone line first: soft tape across the tops of the bones, hair pushed back, the millimetres written down before you look at a single frame. Every row of Table 2 is anchored to that figure.
• Add two of the three stamped numbers — lens width plus bridge width — and subtract your own pupillary distance from the sum. Halve it. Under 3 mm (0.12 in) per eye is comfortable, up to 5 mm (0.20 in) is published as makeable, and past that you are buying the ring patterns in section six.
• Get lens depth in millimetres for anything you mean to order unseen, from the specification rather than the photograph. If nobody will state it, stop there.
• Run Table 2's fourth column against your own width before you weigh any shape advice. If the depth you like lands in its last two rows, that front is going to come up short of your cheekbone whatever the classifier called you, and the three fixes in order of cost are a shallower shape, a heavier end piece, and a smaller frame distance.
• Ask for the frame pupillary distance in writing when you order two candidates and keep whichever clears. The number is on the arm of a pair you already own, and the subtraction takes longer to explain than to do.