What really controls depth of field: aperture, distance, sensor

Depth of field has four inputs, not one: f-number, focus distance, focal length and an agreed blur limit set by sensor and viewing size. Put numbers on all four and the myths about crop sensors and phones sort themselves out.

Love padlocks on a bridge cable in Ljubljana, sharp in the foreground, with city lights behind turned into large soft discs
Photo: Petar Milošević, CC BY-SA 4.0, via Wikimedia Commons

Depth of field is the range of distances that look acceptably sharp, and it is set by four things: the f-number, the focus distance, the focal length, and the circle of confusion, the largest blur spot that still reads as sharp at your final print or screen size. Wider apertures, closer focus and longer lenses make it shallower. For the same framing and the same f-number, a smaller sensor gives deeper focus, which is why a phone at f/1.6 shows almost everything sharp.

Key takeaways
  • Four inputs set depth of field: f-number, focus distance, focal length and circle of confusion.
  • Distance and focal length are squared: halving subject distance does more than opening one stop.
  • 16 cm of sharp depth for a full-frame 50 mm f/1.8 at 2 m; about 2.7 m for a 1-inch phone at f/1.63.
  • Equivalence: APS-C 33 mm f/1.2 behaves like full-frame 50 mm f/1.8.
  • Hyperfocal distance is the focus point that carries sharpness to infinity.

Most explanations stop at "open the aperture for blurry backgrounds". That is one input of four, and it causes the two most common arguments on photography forums:

  • whether crop sensors "have more depth of field";
  • why a phone's f/1.6 lens does not behave like f/1.6 on a full-frame camera.

The answers are arithmetic, so this post does the arithmetic, with a calculator at the end.

The story

On 15 November 1932, an exhibition opened at the M.H. de Young Memorial Museum in San Francisco under a name borrowed from a lens barrel: Group f/64. Seven photographers, including Ansel Adams, Edward Weston and Imogen Cunningham, showed about 80 prints over six weeks. Willard Van Dyke later recalled first suggesting "US 256", an old aperture marking; Adams argued the public would not get it and proposed f/64, one of the smallest stops on a large-format lens.

The choice was a manifesto. At f/64 a view camera holds everything from near rocks to far hills in focus, the opposite of the soft-focus pictorialism the group rejected. The group had dissolved by the end of 1935, but the idea that depth of field is a deliberate decision, not a default, stuck.

Wikipedia: Group f/64 (exhibition, members and the naming)

What exactly determines depth of field?

Four inputs: f-number N, focal length f, focus distance s and circle of confusion c. The thin-lens formulas combine them through the hyperfocal distance, H = f²/(N·c) + f, and then give the near and far limits of acceptable sharpness.

  • Near limit = s(H − f) / (H + s − 2f)
  • Far limit = s(H − f) / (H − s), and infinity when s is at or beyond H
  • Total depth of field = far limit − near limit

These are the standard thin-lens approximations that most depth-of-field calculators use. They assume a symmetric lens and ignore diffraction, but they are accurate enough to predict what you will see.

Every number in this post comes from them, with circle-of-confusion values from the table on Wikipedia's circle-of-confusion article (the d/1500 convention).

16 cmDOF at 2 m: full frame, 50 mm, f/1.8 (my calculation)
2.7 mDOF at 2 m: 1-inch phone sensor, 23 mm equiv., f/1.63
0.029 mmcircle of confusion for full frame (d/1500)
0.011 mmcircle of confusion for a 1-inch sensor

What is the circle of confusion, and who decides it?

The circle of confusion is the largest blur disc on the sensor that a viewer will still see as a point. It is a convention, not a physical constant: the usual assumption is a print about 25 cm wide viewed from 25 cm, on which a 0.2 mm blur spot looks sharp, and that spot is scaled back down to the sensor.

Because the final image is fixed, a smaller sensor must be enlarged more, so its allowed blur on the sensor is smaller. The d/1500 rule divides the sensor diagonal by 1,500.

FormatSensor size (mm)Circle of confusion (mm)Crop factor (diagonal)
Full frame (35 mm)24 × 360.0291.0
APS-C (Nikon, Sony, Pentax)15.7 × 23.60.0191.5
APS-C (Canon)14.8 × 22.20.0181.6
Four Thirds13.5 × 180.0152.0
1-inch type8.8 × 13.20.0112.7
6 × 7 medium format56 × 690.0590.49

Viewing size changes the answer

Change the viewing assumption and depth of field changes with it. Pixel-peeping a 45 MP file at 100% on a monitor is like making a very large print, so the effective c shrinks and the "sharp" zone you perceive gets thinner. That is why a shot that looks crisp on a phone screen shows soft edges when zoomed in.

An Indian lotus bud in sharp focus against a smoothly blurred green background
An Indian lotus bud isolated against a soft background. The background is far beyond the far limit of sharpness, so its detail melts into colour.Photo: Subhrajyoti07, CC BY-SA 4.0, via Wikimedia Commons

How much does each input change depth of field?

Roughly, total depth of field scales with N·c·s²/f² when the subject is well inside the hyperfocal distance:

Double this inputDepth of field becomesWhy
Focus distance sabout 4×squared in the numerator
Focal length fabout ¼squared in the denominator
f-number Nabout 2×linear
Circle of confusion cabout 2×linear

Distance and focal length matter most because they are squared.

Halving your distance to the subject does more for background blur than opening the lens by a full stop.
Total depth of field with the subject at 2 m
See the data as a table
Depth of field (cm)
Full frame, 85 mm, f/1.85.5
Full frame, 50 mm, f/1.816.3
APS-C, 33 mm, f/1.216.2
APS-C, 33 mm, f/1.824.4
Full frame, 50 mm, f/874.8
1-inch phone, 23 mm equiv., f/1.63265.7
Full frame, 24 mm, f/8868.5

Source: My calculation with thin-lens DOF formulas and Wikipedia circle-of-confusion values (0.029 mm full frame, 0.019 mm APS-C, 0.011 mm 1-inch) · Phone row assumes an actual focal length of 8.5 mm (23 mm divided by a 2.72 crop factor) as on the Xiaomi 15 Ultra main camera.

The 85 mm portrait lens at f/1.8 gives a sharp zone of about 5.5 cm at 2 m, enough for the eyes but not the ears. Stop the 50 mm down from f/1.8 to f/8 and the zone grows from 16 cm to 75 cm.

Calculated in this post from the thin-lens formulas above, with these example inputs (85 mm, f/1.8, 2 m) and the 0.029 mm full-frame circle of confusion from Wikipedia: Circle of confusion.

Full frame, 50 mm, f/8, focused at 2 m camera 2.0 m1.69 m2.44 m 31 cm in front · 44 cm behind blurredblurred Scale: 1 m ≈ 155 px from the camera.
The sharp zone from the calculator, drawn to scale. The zone is lopsided, deeper behind the focus point; move focus back and the far edge races away until it reaches infinity at the hyperfocal distance.

A 24 mm lens at f/8 focused at 2 m is sharp from 1.1 m to 9.8 m. For the related question of why longer lenses also seem to "compress" backgrounds, see focal length and perspective compression.

Do crop sensors have more depth of field?

Only when you compare like with like in framing. At the same f-number and the same framing, a smaller sensor gives deeper focus by about the crop factor, because it needs a shorter lens to frame the shot. At the same lens, f-number and distance, the smaller sensor actually gives slightly shallower depth of field, because its stricter circle of confusion outweighs everything else, and it shows a tighter crop.

The equivalence rule

That is the nuance behind the Stack Exchange thread on whether crop factor can be used to compute the change exactly. The rule that works is equivalence: multiply both focal length and f-number by the crop factor to find the full-frame setting with the same framing and the same depth of field.

An APS-C camera with a 33 mm f/1.2 lens behaves like a full-frame 50 mm f/1.8: my calculation gives 16.2 cm and 16.3 cm of depth at 2 m. On an APS-C body like the Sony ZV-E10, the 1.5 crop means you need about one and a bit stops wider aperture to match the background blur of full frame.

SetupFocal lengthf-numberFull-frame equivalentDOF at 2 m
Full frame50 mmf/1.850 mm f/1.816.3 cm
APS-C (1.5×)33 mmf/1.850 mm f/2.724.4 cm
APS-C (1.5×)33 mmf/1.250 mm f/1.816.2 cm
1-inch phone (2.72×)8.5 mmf/1.6323 mm f/4.42.66 m

Is crop factor a bad thing?

As the third forum thread asks: not for depth. deeper focus at the same f-number is welcome for landscapes, product shots and street work. It costs light, though: equivalence applies to total light too, so a smaller sensor at the same f-number and shutter speed collects less total light and shows more noise. The exposure side of that trade is in the exposure triangle in stops.

A lit taxi sign on a car roof at night in focus, with city lights behind it turned into large soft discs
Out-of-focus highlights render as discs whose size grows with the aperture's physical diameter. This is what "bokeh" describes, and what phone portrait modes imitate.Photo: Petar Milošević, CC BY-SA 4.0, via Wikimedia Commons

Why does a phone at f/1.6 still have deep focus?

Because the f-number is a ratio, and the phone's real focal length is tiny. The Xiaomi 15 Ultra's main camera, one of the largest phone sensors sold in 2025, is a 1-inch type at f/1.63 with a 23 mm equivalent field of view, which works out to an actual focal length of about 8.5 mm. Its entrance pupil is only about 5.2 mm across, against 27.8 mm for a 50 mm f/1.8 lens.

Run the numbers: focused at 2 m, that phone camera is sharp from about 1.33 m to 3.99 m, and its hyperfocal distance is only about 4 m, so focusing there makes everything from 2 m to infinity look sharp. Most phones use sensors smaller than 1 inch, which makes the zone deeper still. Optically, a phone cannot blur a background behind a person at normal distances; the blur in portrait mode is computed.

Synthetic Depth-of-Field with a Single-Camera Mobile PhoneNeal Wadhwa, Rahul Garg, David E. Jacobs et al., with Marc Levoy · ACM Transactions on Graphics (SIGGRAPH) · 2018

The Google Research team behind Pixel portrait mode starts from the problem stated above: phone cameras' short focal lengths and small apertures capture "nearly all-in-focus images". Their system combines a neural network that segments people with a depth map from the sensor's dual-pixel autofocus, which works like a tiny stereo pair with a baseline of about 1 mm.

It renders disc-shaped blur proportional to disparity, deliberately keeps a band of depths around the subject sharp so novices do not miss focus, and adds synthetic noise back into blurred areas to avoid a cardboard-cutout look. A 5.4 MP result took about 4 seconds on the phone, after a forced 1.5× crop that also pushes the photographer back from the face. The limit: with a 1 mm baseline, depth from dual pixels alone only works for close subjects, so scenes without people rely on much weaker cues.

Read the paper (free)

I go deeper into burst merging, depth maps and HDR in how phone cameras use computational photography.

Does blur change how big a scene looks?

Yes. Our visual system reads a steep blur gradient as a sign that the scene is close and small, because only nearby objects produce shallow depth of field in the eye. That is the trick behind tilt-shift miniature photos, and it means depth of field is not just about isolating subjects; it also tells viewers about scale.

Using Blur to Affect Perceived Distance and SizeRobin T. Held, Emily A. Cooper, James F. O'Brien, Martin S. Banks · ACM Transactions on Graphics · 2010

The authors built a probabilistic model of how viewers combine defocus blur with other depth cues to estimate absolute distance, and used it to add blur to sharp images so that scenes appeared closer and smaller. In an experiment, viewers' distance estimates for images with physically correct blur matched the model's predictions closely, and so did their responses to common approximations: some shortcuts produced the miniature effect, others did not, as predicted. The limit: a 2023 study (below) notes that the numerical distance-estimation task was error-prone and a share of participants had to be excluded.

Read the paper (free)

Meese, Baker and Summers tested the same idea with a forced-choice task in PLOS ONE (2023): 108 participants had to pick which of a pair of photos showed a full-size railway scene and which a 1:76 scale model. Adding a fake blur gradient to full-size scenes made people pick them as the model significantly more often, and the orientation of the gradient relative to the ground mattered more than its steepness.

sensor in-focus point far point blur circle on the sensor aperture
A point outside the focus distance converges before (or after) the sensor and lands as a disc. When that disc is smaller than the circle of confusion, the point still looks sharp. A narrower aperture narrows the cone of rays, so the disc shrinks.

What is hyperfocal distance and how do you use it?

The hyperfocal distance is the closest focus setting at which the far limit of sharpness reaches infinity. Focus there and everything from half that distance to the horizon looks sharp. It is the landscape and street photographer's shortcut, and it is the reason phones rarely need to refocus for distant scenes.

Camera and lensf-numberHyperfocal distanceSharp from (when focused at H)
APS-C, 16 mmf/81.70 m0.85 m to infinity
Full frame, 24 mmf/82.51 m1.26 m to infinity
1-inch phone, 8.5 mm (23 mm equiv.)f/1.634.0 m2.0 m to infinity
Full frame, 35 mmf/85.32 m2.66 m to infinity
Full frame, 50 mmf/810.8 m5.4 m to infinity

Reading a lens's depth-of-field scale

Old manual lenses printed this on the barrel as a depth-of-field scale: pairs of f-number marks either side of the focus index.

  1. Pick the aperture. Find the pair of marks for, say, f/8.
  2. Set infinity. Turn the focus ring until ∞ sits against the right-hand f/8 mark.
  3. Read the near limit. The distance opposite the left-hand f/8 mark is where sharpness starts.

Remember the scale assumes the manufacturer's circle of confusion, usually the generous 0.03 mm, so for large prints stop down one more than the scale suggests.

The barrel of a Canon FD 70-210 mm zoom lens with coloured depth-of-field lines for different apertures beside the focus scale
A depth-of-field scale on a Canon FD 70-210 mm zoom. The coloured lines fan out because depth of field shrinks as focal length grows, the f-squared term in the formula.Photo: Karl Thomas Moore, CC BY-SA 4.0, via Wikimedia Commons

Depth of field calculator

Use the real focal length of the lens (not the equivalent) and the circle of confusion for your sensor from the table above: 0.029 for full frame, 0.019 for most APS-C, 0.015 for Four Thirds, 0.011 for 1-inch.

Depth of field and hyperfocal distance

The standard lens in all these examples, 50 mm on full frame, earned that status for reasons beyond depth of field; that history is in why 50 mm is the standard lens.

Questions people ask

Does a wider aperture always mean a blurrier background?

At the same focal length, distance and sensor, yes. Across different cameras, compare the physical aperture diameter (focal length divided by f-number) instead: a 50 mm f/1.8 has a 27.8 mm pupil, a phone's 8.5 mm f/1.63 only 5.2 mm, which is why the phone's background stays sharp.

Is depth of field split one-third in front, two-thirds behind?

Only at one particular distance. Close up, the split approaches half and half; as you approach the hyperfocal distance, the far side grows towards infinity. In my 50 mm f/8 example at 2 m, the zone runs 31 cm in front and 44 cm behind.

Why does my photo look sharp on the phone but soft on a laptop?

Because depth of field depends on viewing size. A small phone screen tolerates a larger blur spot; a big monitor or 100% crop enlarges the image and shrinks the allowed circle of confusion, so parts you thought were sharp fall outside the zone.

Does stopping down to f/22 give the most sharpness?

It gives the most depth, but diffraction softens the whole image as the aperture shrinks. As a working rule of thumb, many full-frame and APS-C lenses are sharpest overall around f/5.6 to f/11; beyond that you trade crispness for depth, and the exact point depends on the lens and pixel size.

Sources

  1. Wikipedia: Group f/64 (1932 exhibition and naming)
  2. Wikipedia: Circle of confusion (CoC values by format, d/1500 convention, 0.2 mm at 25 cm viewing)
  3. Cambridge in Colour: Depth of field
  4. Wadhwa et al.: Synthetic Depth-of-Field with a Single-Camera Mobile Phone, ACM TOG / SIGGRAPH (2018)
  5. Held, Cooper, O'Brien and Banks: Using Blur to Affect Perceived Distance and Size, ACM TOG (2010)
  6. Meese, Baker and Summers: Blurring the boundary between models and reality: Visual perception of scale assessed by performance, PLOS ONE (2023)
  7. Digital Camera World: Xiaomi 15 Ultra review (2025; 1-inch LYT-900, 23 mm f/1.63)
  8. Notebookcheck: Xiaomi 17 Ultra and its 1-inch sensor (2025)
  9. DXOMARK: Xiaomi 17 Ultra camera test (2026)
  10. Photography Stack Exchange: What exactly determines depth of field?
  11. Photography Stack Exchange: Can a smaller sensor's crop factor be used to calculate the exact increase in depth of field?
  12. Photography Stack Exchange: Is crop factor a bad thing?
depth of fieldcircle of confusionhyperfocal distanceaperturecrop factorequivalencebokehportrait mode
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Artblog.

Keep reading in Artblog

From the other blogs