Why your depth of field calculator disagrees with your photographs

Here is a situation that has made a lot of photographers distrust the maths.

You are shooting a landscape. You put a 24mm lens on a full frame body, set f/11, focus about four metres in, and a depth of field calculator tells you everything from roughly 1.7 metres to infinity will be sharp. You take the picture. At home you open it, zoom to 100%, look at the foreground rocks at two metres, and they are clearly not sharp. Not disastrous, but not sharp.

The calculator was not wrong. It answered a question you did not realise you were asking.

The number in the middle of every depth of field formula

Only one plane is ever in true focus. Everything in front of and behind it is a blur circle, growing steadily with distance from that plane. There is no boundary in the physics where sharp becomes unsharp.

Depth of field exists because we impose a boundary: pick a blur circle small enough to be invisible, and the range where blur stays under that size becomes "the sharp bit". That threshold is the circle of confusion, and every depth of field number you have ever read depends entirely on which value someone picked.

The conventional choice is the sensor diagonal divided by about 1500. For full frame that is 43.27mm / 1500, or roughly 0.029mm — the 0.03mm you see quoted everywhere. Smaller formats get proportionally smaller values.

Where does 1500 come from? It works backwards from a person. Take an 8x10 inch print, hold it at about 25cm, assume normal vision resolving around five line pairs per millimetre at that distance, and work out how large a blur spot on the negative can be before it becomes visible in the print. That is the number.

So the honest full statement of what a calculator tells you is: given this focal length, aperture and distance, here is the range that will look sharp in an 8x10 print viewed at arm's length by someone with average eyesight.

Nobody says that. They say "here is the depth of field", and the assumption disappears.

What 100% on a monitor actually is

Now consider what you did when you zoomed to 100%.

One image pixel became one screen pixel. On a 24 megapixel file at a typical monitor pitch, that is equivalent to standing 40cm away from a print about 1.5 metres wide. On a 45 megapixel file it is a print over two metres wide.

You did not check whether the image met the 8x10 standard. You checked whether it met a standard five or six times stricter, then concluded the calculator lied.

Both results are correct. The rocks are sharp for an 8x10 print and not sharp for a two metre one. This is not a fudge. It is what depth of field has always meant, and the reason the concept feels slippery is that it was defined in an era when the output size was known in advance and it no longer is.

Setting the circle of confusion to match your output

The fix is to pick a value that matches what you actually do with the photographs, rather than what a 1930s print convention assumed.

For pixel-level sharpness, use roughly 1.5 to 2 times your sensor's pixel pitch. Divide the sensor width in millimetres by the horizontal pixel count: a 45MP full frame body is 36mm / 8256, about 4.4µm, so a pixel-peeping circle of confusion is around 0.007 to 0.009mm. That is three to four times stricter than the standard 0.03mm.

The consequences follow directly from the formulas. Depth of field shrinks by roughly the same factor. Hyperfocal distance, which is f² / (N × c), scales inversely with c — so making the circle of confusion four times smaller makes the hyperfocal distance four times longer. A 24mm f/11 lens with a hyperfocal distance of about 3.5 metres by the traditional number moves out past 14 metres by the strict one, and the entire near half of your composition falls outside it.

That single relationship explains most of the arguments about hyperfocal focusing on the internet. The people saying it works and the people saying it produces mush are using different values of c and neither side mentions it.

For large prints, something in the region of 0.015 to 0.02mm on full frame is realistic. For web and small prints, the traditional 0.03mm is if anything conservative, since a photograph viewed on a phone is being held to a far looser standard than an 8x10 print.

The depth of field calculator here uses the conventional diagonal/1500 figure and says so on the page, along with the formula it evaluates, so you can see exactly what assumption produced the answer.

The other places the model bends

The circle of confusion accounts for most of the disagreement between calculator and photograph. A few other things account for the rest.

Close focus. Standard depth of field formulas use a thin lens approximation with the focus distance measured from the front principal plane, and they degrade as you approach 1:1 magnification. Closer than about twice the focal length they stop describing reality. In macro work, use magnification-based formulas instead — depth of field near 1:1 is a function of magnification and effective aperture, not of subject distance, which is why the magnification calculator is the more useful tool down there.

Focus breathing. Many lenses change focal length slightly as they focus, particularly internally focusing designs. A "70-200mm" at close range may be closer to 150mm. Feed the marked focal length into a calculator and the answer is right for a focal length the lens is not currently at.

Field curvature. Depth of field maths assumes the plane of focus is a plane. On real lenses it is usually a shallow dish. Corners can be focused a measurable distance nearer or further than the centre, which is why the corners of a landscape often disagree with the calculator even when the middle agrees perfectly.

The one-third rule. The old idea that depth of field extends one-third in front of the focus point and two-thirds behind is only approximately true at moderate distances. Close up the split approaches even. Approaching the hyperfocal distance the rear extends towards infinity and the ratio becomes meaningless. It is a rule of thumb that was never meant to survive being quoted as a law.

Diffraction. Depth of field formulas assume that stopping down only ever adds sharpness at the edges of the zone. Past a certain aperture, diffraction is softening the whole frame including the plane of focus, so the calculator promises more sharp depth while the image gets worse overall. That trade is covered in the diffraction limit is not a cliff.

How to use the numbers without being misled

Decide what the photograph is for before you decide where to focus. A frame destined for a gallery print and a frame destined for a website genuinely have different depth of field, because depth of field is a statement about a viewer, not only about a lens.

If you shoot high resolution files and inspect them closely, treat published depth of field figures as optimistic by a factor of three or so, or set a circle of confusion that matches your pixel pitch and work from that.

Where it matters and you can afford the frames, focus bracket. Three shots at different focus distances is a more reliable answer than any calculation, and the focus stacking calculator will tell you how many you need.

And when a calculator and a photograph disagree, the useful question is not which one is lying. It is which viewing standard each of them is using, because that is almost always the whole of the difference.