Circle of Confusion, Explained

What the circle of confusion is

A point in front of or behind the focus plane doesn't image as a point. It lands on the sensor as a small disc, and the disc grows the farther the point is from focus. The circle of confusion (CoC) is the largest disc you are willing to call sharp: small enough that, after the picture is enlarged to its viewing size, the eye still reads it as a point.

It is a choice, not a property of the lens. Every depth of field and hyperfocal figure depends on it, which is why two calculators given the same lens, stop and distance can disagree: they are being strict to different degrees. See how to calculate depth of field for the formulas it feeds.

Where the numbers come from

Why 1,500: viewed from a distance about equal to the picture's diagonal, a blur of 1/1,500 of the diagonal subtends about two minutes of arc, which the eye takes for a point. ZEISS puts the strictest sensible standard at twice as strict (half the size), diagonal ÷ 3,000, and notes that 1950s practice was much looser: 0.05mm for 35mm, about 1/865 of the diagonal.

Motion-picture practice settled on 0.025mm (0.001 inch) for 35mm. That is the value behind cine lens makers' depth of field tables, Cooke's among them. For a Super 35 sensor it is slightly more lenient than the diagonal rule, which gives about 0.023mm. Enter either value as a custom circle of confusion in the free depth of field calculator to see the difference it makes.

Circle of confusion by format

How it changes depth of field

Depth of field scales almost in proportion to the circle of confusion, and the hyperfocal distance in inverse proportion. Same lens, same stop, same distance:

The format matters too, but less than it seems. On a bigger sensor the same lens gets a larger circle of confusion and more depth; you then fit a longer lens to hold the framing, which takes more away. The sensor size section of the depth of field guide works through Super 35 against full frame.

Choosing a tighter value

The standard values assume the whole frame is seen at a normal size. Plan with a smaller circle of confusion when that isn't true:

• Cropping in post. A 1.5× punch-in enlarges blur by 1.5×, so divide the circle of confusion by 1.5: 0.025mm becomes about 0.017mm.
• Large screens. The front rows of a theater see the picture at a far wider angle than a viewer at a distance equal to its diagonal. Planning with half the standard value is the cautious choice for big-screen work.
• Visual effects plates. Tracking and keying see softness the eye forgives. Ask the VFX supervisor what they need before the day.
• High resolution. A 0.030mm blur on an ALEXA Mini LF spans about 3.6 photosites (8.25µm each). Picture that is "in focus" by the standard is not pixel-sharp, and on a large display it can show.

The trade-off is real: half the circle of confusion halves your depth of field. Pick the value in prep with the DP, put it on the camera report, and use the same one for every calculation on the job.

Diffraction: the other blur

Stopping down adds depth of field but also adds diffraction blur. Even a perfectly focused point images as an Airy disk, about 2.44 × wavelength × f-number across. For green light (550nm) that is about 1.34 × the f-number in micrometers, and at small stops it rivals the circle of confusion itself.

Past about f/16 on Super 35, diffraction alone is close to the whole blur budget, so stopping further down makes everything a little softer even as the depth of field figure grows. CineMath's hyperfocal mode warns when the stop is past f/16 for this reason. When a split needs T22 or more, look for another way to hold both subjects.

Questions

Next step: decide your circle of confusion for the job, then use it everywhere. CineMath's Depth of Field calculator has a preset for each format and takes a custom value, so every limit, hyperfocal distance and split uses the same standard. The free web version has the same presets.