← Back to DoF Viewer

Expert mode: understanding the circle of confusion

A little theory, in pictures.

The circle of confusion (CoC) is the size of the smallest spot the eye sees as sharp. It depends on three things: the viewer's visual acuity, the viewing distance, and the enlargement ratio between the sensor and the final picture. The usual 0.03 mm value for 24×36 assumes one very specific case. Here is where it comes from, why it is not always enough, and how Expert mode recalculates it for a more reliable depth of field (DoF).

The standard viewing conditions

The classic calculation assumes the picture is viewed from 25 cm. Our field of view is about 60°: at that distance it covers a 29 cm circle, in which a 16 × 24 cm picture fits.

Standard viewing conditionsTop view: an eye 25 cm from a picture, within a 60° field of view. Front view: a 29 cm circle with a 16 × 24 cm rectangle inscribed in it.Top viewFront view60°Picture16 × 24 cmEye25 cm24 cm16 cm29 cm
Figure 1 — At 25 cm, a 60° field of view covers a 29 cm circle, i.e. a 16 × 24 cm picture.

The acuity of the eye

At 25 cm, the eye resolves details of about 0.1 mm. That is equivalent to a test chart of 5 line pairs per millimetre (lp/mm). One line pair measures 0.2 mm: that is the CoC on the final picture.

The eye and the stripesA test chart of black and white stripes, 5 pairs per millimetre. Two red circles are one stripe pair wide, i.e. 0.2 mm; one stripe is 0.1 mm.1 mm = 5 lp/mm0.1 mm0.2 mm = CoCCoC
Figure 2 — 0.1 mm at 25 cm: 5 lp/mm, i.e. a 0.2 mm CoC on the final picture.

From print to sensor

To find what this CoC means on the sensor, apply the enlargement factor. From 24 × 36 mm to 240 × 160 mm, it is about 6.7. The CoC on the sensor is therefore 0.2 / 6.7 ≈ 0.03 mm.

Carrying it over to the sensorA 240 × 160 mm picture at 250 mm, linked by dotted lines to a 24 × 36 mm sensor. The 0.2 mm CoC on the picture becomes 0.03 mm on the sensor (divided by 6.7).Picture 240 × 160 mm0.2 mm0.03 mmSensor 24 × 36 mm÷ 6.7250 mm
Figure 3 — CoC on the sensor = 0.2 mm / 6.7 ≈ 0.03 mm.

But only under those conditions

The 0.03 mm CoC holds for a 16 × 24 cm picture viewed from 25 cm, with 0.1 mm acuity. What happens with a 50 × 75 cm print, or when viewing from 50 cm?

When the assumption no longer holdsTwo top views. A: a 50 × 75 cm print viewed from 25 cm, much larger than the reference. B: a 16 × 24 cm picture viewed from 50 cm. In both cases the CoC is a question mark.50 × 75 cm printEye25 cm?CoC16 × 24 cmViewed from 50 cmEye50 cm?CoC16 × 24 cm
Figure 4 — If the print size or the distance changes, 0.03 mm is no longer the right criterion.
The 0.03 mm CoC is no longer accurate enough, and neither is the DoF calculation. That is what Expert mode is for.

The three Expert mode settings

Expert mode keeps the 16 × 24 cm picture as its reference, but recalculates the CoC from the real conditions: visual acuity, viewing distance and print size.

Visual acuity

Three levels are offered. Each sets the smallest visible detail as a fraction of the viewing distance.

The three acuity levelsThree test charts to scale: High (0.083 mm at 25 cm, 1/3000 of the distance), Medium (0.1 mm, 1/2500), Low (0.2 mm, 1/1250). High has the finest stripes, Low the widest.LevelTest chart to scaleVisible detailProportionHigh0.083 mm at 25 cm1/3000 of the distanceMedium0.1 mm at 25 cm1/2500 of the distanceLow0.2 mm at 25 cm1/1250 of the distance
Figure 5 — The three acuity levels, drawn to scale: the lower the acuity, the larger the smallest visible detail.

Viewing distance

The further away you are, the fewer details the eye resolves. The smallest visible detail is proportional to distance: at 50 cm it is twice as large as at 25 cm.

Distance and acuityLog-log chart: the smallest visible detail grows in proportion to viewing distance (25 cm to 5 m), for High, Medium and Low.0.10.20.512525 cm50 cm1 m2 m5 mVisible detail (mm)Viewing distanceLow 4 mmMedium 2 mmHigh 1.7 mm
Figure 6 — Smallest visible detail against distance (log scales). The nominal CoC is twice that detail.

Print size

A new enlargement ratio is computed to keep the CoC consistent with the chosen acuity and distance. The 16 × 24 cm reference is in orange.

Print sizesNested rectangles to scale: 10×15, 16×24 (reference, in orange), 20×30, 50×75 and 76×114 cm, with their ratio R to the reference.76 × 114R = 4.7550 × 75R = 3.120 × 30R = 1.2516 × 24ref.10 × 15R = 0.63
Figure 7 — Available print sizes, to scale. R = ratio to the 16 × 24 cm reference picture.
The final CoC (Expert CoC) takes all three conditions into account.

The reasoning, step by step

How do these three settings lead to the final CoC? In six steps, with the same test chart and the same red circles. Enlargements are drawn with a ratio of 2, for illustration.

The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.1ReferenceNominal CoC
1. The reference CoC matches the chosen acuity and distance.
The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.2Enlargement× 2 (example)
2. When the picture grows, the stripes grow too.
The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.3MismatchNominal CoC unchanged≠
3. But the eye has stayed at the same distance: acuity and CoC no longer match.
The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.4New CoCNominal CoC / enlargement ratio
4. A new CoC is worked out that takes the enlargement into account.
The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.5Enlargement× 2 (example)
5. The picture is enlarged with that ratio.
The reasoning in 6 stepsSix thumbnails: 1 reference, 2 enlargement, 3 mismatch between CoC and stripes, 4 new CoC, 5 enlargement with the new CoC, 6 alignment.6AlignmentCoC and stripes aligned
6. Stripes and new CoC are consistent again with the chosen conditions: it can be used to compute the DoF.

Last step: back to the sensor

The CoC still has to be carried over to the sensor, using the ratio between the reference picture and the sensor.

Back to the sensorThe enlarged picture with its CoC, linked by dotted lines to a small sensor where the CoC is divided by R_sensor.Enlarged pictureCoCnominal​ / Renlargement​Sensor÷ Rsensor​(CoCnominal​ / Renlargement​) / Rsensor​
Figure 9 — The CoC on the sensor is (CoCnominal / Renlargement) / Rsensor.

In short

Four settings go into the Expert CoC. The result is a more accurate estimate of the DoF, because it takes the viewing conditions of the final picture into account.

Summary diagramFour inputs (visual acuity, viewing distance, print size, sensor format) lead to the Expert CoC calculation, then to a more reliable depth of field.Visual acuityViewing distancePrint sizeSensor formatExpert CoC calculationMore reliabledepth of field
Figure 10 — From the user's settings to the DoF: inputs in blue, result in orange.

CoCnominal = 2 × acuity(d)

with acuity(d) = d / 2500 (Medium), d / 3000 (High) or d / 1250 (Low)

CoCexpert = CoCnominal / (Renlargement × Rsensor)

  • Rsensor: ratio between the 16 × 24 cm reference picture and the sensor (≈ 6.7 for 24×36).
  • Renlargement: ratio between the chosen print and the 16 × 24 cm reference picture.
  • The factor 2 comes from acuity: 0.1 mm corresponds to a 0.2 mm CoC, i.e. one line pair.

In other words: CoC on the sensor = smallest visible detail on the print ÷ total enlargement (print / sensor).

Worked examples (24×36, Medium acuity)

PrintDistanceCoC on the printTotal enlargementCoC on the sensor
16 × 24 cm25 cm0.2 mm≈ 6.7×≈ 0.030 mm
50 × 75 cm25 cm0.2 mm≈ 20.8×≈ 0.0096 mm
50 × 75 cm50 cm0.4 mm≈ 20.8×≈ 0.019 mm
16 × 24 cm50 cm0.4 mm≈ 6.7×≈ 0.060 mm

A large print viewed up close calls for a much smaller CoC, and therefore less depth of field; viewing from further away relaxes it.

In the app

In Expert mode, the CoC can become smaller than the sensor's pixel size. DoF Viewer still computes the DoF, but the value is no longer realistic; the fields concerned turn red.

Back to Expert mode