Reciprocity and Reciprocity Failure

Reciprocity is the rule that exposure equals the brightness of the light multiplied by the time it falls on the film or sensor, so you can halve one, double the other and get the same picture brightness. Reciprocity failure is film breaking that rule: at long exposures, and at extremely short ones, film responds less than the arithmetic predicts and needs more exposure than the meter says.

The first half is why every pair of equivalent settings works, on any camera. The second half matters to anyone who shoots film at night, through a dense ND filter or through a pinhole, because a metered 30 seconds can need a minute and a half, and the gap grows faster the longer you go. Digital sensors do not suffer from it in any practical way.

The reciprocity law: light times time

Exposure is the amount of light that reaches each part of the film or sensor, and it is the product of two things: how bright the light is there and how long it lasts. The rule is also called the Bunsen and Roscoe law, after the two chemists who described it for light-sensitive materials.

Each full stop of aperture halves or doubles the light, and each full stop of shutter speed halves or doubles the time. Trade one stop of one for one stop of the other and the product does not change. All of these give the same exposure:

The brightness of the picture is the same in every row. What changes is depth of field and how motion is recorded, which is why equivalent exposures are equal in brightness but not identical pictures. Each row is the same exposure value, and that single number is only meaningful because reciprocity holds.

ISO sits in the same chain, but it is neither light nor time. It sets how much exposure the meter assumes you need, so doubling the ISO lets you halve the light or halve the time and still reach the same picture brightness.

Lock the exposure, move one setting, and watch the other pay for it. Every combination the tool shows lets the same light through.

Not sure why a stop is a doubling, or how a full stop differs from a third? Stops in photography builds the scale every row above is counted on.

What reciprocity failure is

Film obeys the law only across a middle band of exposure times. For many films that band runs from around 1/1000 second to about a second, but films differ a great deal: some need about a stop of extra exposure at a metered one second, while some modern fine-grain films need no correction until about two minutes. Outside its band, a film gives less density than the light times time product predicts. It behaves as if it had lost speed.

The practical result: give film the time the meter suggests for a long exposure and the negative or slide comes out underexposed, worst in the shadows.

The astronomer Karl Schwarzschild measured the effect on photographic plates and described it with an exponent: the effect of the light behaves like intensity times time raised to a power a little below one, rather than intensity times time. That exponent is why the shortfall is not a fixed number of stops. It grows the longer the exposure runs.

Why film loses speed in dim light

Film records light in tiny crystals of silver halide. Each photon a crystal absorbs frees an electron, and that electron turns one silver ion into a silver atom. A crystal becomes developable once a few silver atoms gather into a stable speck. A single atom on its own is unstable and can break apart again within a short time.

In bright light, photons arrive close together, so a crystal builds a stable speck before its first atom can fall apart. In dim light, they arrive so far apart that many lone atoms break up before the next photon reaches them. That light does nothing. The dimmer the light, the more of it is wasted, which is why the film seems slower in exactly the conditions where you need every photon.

The same mechanism has a cousin, the intermittency effect: on film, many short exposures that add up to one second do not give the same density as a single one-second exposure, because each gap gives lone atoms time to break up. It is one reason a multiple exposure built from several dim exposures can come out thinner on film than the arithmetic says.

Failure at extremely short exposures

Very brief, very intense light causes the opposite problem. Electrons are freed so quickly that they start many small specks in the same crystal instead of a few that grow large enough to develop. Density again falls short, and color film can shift. It shows up at durations of roughly 1/10,000 second and shorter, such as a small flash fired at close range on its lowest power. Ordinary flash and shutter speeds are well inside the safe band, so most photographers never meet it.

How much extra exposure film needs

Film makers publish reciprocity data for each emulsion, either as a table of metered time against corrected time, or as a formula. Many black and white data sheets use a power formula:

Corrected time = metered time^p, with both times in seconds.

The exponent p is a number a little above 1, fitted by the maker to its own tests of that film. Published values for black and white films commonly fall somewhere between about 1.25 and 1.45. Because 1 raised to any power is 1, the formula leaves a one-second exposure alone and only bites above that. The table shows three illustrative exponents across that range. Use the number for your own film, not these.

Two things stand out. The correction keeps growing: with p = 1.31, the extra exposure is 0.3 stop at 2 seconds and 2.1 stops at 2 minutes. And the spread between films is large: at a metered 2 minutes, the most forgiving exponent here asks for under 7 minutes and the most demanding for over 17.

The formula is a fitted curve, not a law of physics. It drifts at very long times, and data sheets say so. For anything past a few minutes, bracket: shoot the corrected time and one frame with a stop more.

Open the aperture or add time

There are two ways to give film the exposure it is missing, and the choice matters more than it looks. You can keep the aperture and lengthen the time, or open the aperture and accept a shorter, less corrected time.

Opening up is more powerful than it seems, because it shortens the metered time before the exponent compounds it. With p = 1.31, one stop of aperture shortens the corrected time by 1.31 stops, a factor of about 2.5, not 2:

The cost is depth of field, and sometimes the look you wanted: a shorter time smooths water and clouds less. When the near-to-far sharpness matters, stay at the small aperture and take the time. When it does not, opening a stop or two can turn a nine-minute wait into three and a half. Some film makers’ tables list both choices side by side for this reason.

Correcting a long exposure on film, step by step

1. Meter the scene and work out the time at the aperture and ISO you have chosen. A handheld light meter helps in dim light where the camera’s meter runs out of range.
2. If an ND filter is on the lens, add its stops first. A 10-stop filter multiplies the time by 1,024, so a 1/30 second reading becomes about 34 seconds.
3. Apply the film’s reciprocity data to that total. With p = 1.31, 34 seconds becomes about 102 seconds, roughly 1 minute 42.
4. If the corrected time is longer than you can live with, open the aperture and repeat steps 1 to 3.
5. Set bulb mode, lock a cable release, and time the exposure with a watch or timer.
6. Write down the metered and the corrected time for every frame. A notebook of results for your own film and developer is worth more than any general table.

Reciprocity failure can also help. Because a film’s long exposures run longer than the meter says, a 6-stop ND on film can give the drawn-out water and clouds that would need a stronger filter on digital.

Planning the shot itself, from the tripod to the timing? Long exposure photography covers filters, timing and keeping the camera still for minutes at a time.

Contrast and color shifts on long film exposures

Contrast rises

Within one exposure, the dim shadows are in dimmer light than the highlights, so they suffer more reciprocity failure. The highlights build density close to normally while the shadows fall behind, and the negative comes out more contrasty than the scene. That is why many black and white data sheets pair the exposure correction with shorter development for very long exposures: less development holds the highlights down while the extra exposure lifts the shadows. Follow the data sheet for your film and developer rather than guessing.

Want to control contrast through exposure and development on purpose? The Zone System is built on exactly that pairing.

Color drifts

Color film has three light-sensitive layers, and each loses speed at its own rate on a long exposure. The balance drifts, often toward green, blue or magenta depending on the film. Data sheets for color films may list a color correction filter alongside the exposure correction for long times. Slide film shows the shift most plainly, because there is no printing or scanning step in which to correct it. Some photographers welcome the shift as part of the look of night on film.

Where you will meet reciprocity failure

• Night scenes on film. City lights, moonlit landscapes and night photography generally run to metered times of many seconds or minutes.
• Stars on film. Star trails and astrophotography on film involve exposures of minutes to hours. Faint stars suffered so badly that astronomers once chilled or chemically treated their plates to reduce the effect.
• Pinhole cameras. Pinholes work at around f/150 to f/300, so even a sunlit scene can need seconds and an interior many minutes.
• Daytime long exposures with dense ND filters on film, to smooth water and clouds.
• Close-up and large format work, where bellows extension and small apertures push times past a second.

Choosing a film for night or long exposure work? Film stocks compared sets the main types side by side, so you know what to look up on the data sheet.

Does reciprocity failure affect digital cameras?

No, not in any way you need to correct. A digital sensor counts the light that reaches each pixel in a very nearly linear way, so twice the time gives twice the signal at 1/4000 second or at 4 minutes. The time the meter gives is the time to use, and a stop is always exactly a doubling.

Long digital exposures have different problems:

• Heat and noise. The sensor warms during long exposures, which adds thermal noise, hot pixels and sometimes a glow along one edge. Higher ISO makes the luminance noise worse.
• Long exposure noise reduction. Many cameras follow the real exposure with a dark frame of the same length and subtract it to remove hot pixels. A 2-minute exposure then ties up the camera for about 4.
• Battery drain, especially in the cold.

At very short durations, an electronic shutter can show banding under flickering lights, but that comes from the sensor being read out line by line, not from reciprocity failure.

Seeing bright colored specks in your long digital exposures? Hot, stuck and dead pixels explains which is which and how to remove them.

Reciprocity is not the reciprocal rule

The names sound alike and get mixed up. Reciprocity is about exposure: light times time. The reciprocal rule is a guide to handheld sharpness: use a shutter speed of at least 1 over the focal length, so 1/100 second or faster with a 100mm lens on a full frame camera. One keeps brightness right; the other keeps camera shake down.

Common mistakes with reciprocity

• Treating the correction as a fixed number of stops. Fix: it grows with time. With p = 1.31 it is 0.6 stop at 4 seconds and 2.1 stops at 2 minutes, so work it out for each exposure.
• Applying the correction before the filter factor. Fix: add the ND filter’s stops first, then correct the total time. Correcting 1/30 second does nothing; correcting 34 seconds adds over a minute.
• Thinking failure overexposes. Fix: it underexposes, and the shadows most of all, which is why contrast rises.
• Using one film’s numbers for another. Fix: films range from needing nothing for two minutes to needing a stop at one second. Use the data for the exact film.
• Adding reciprocity corrections on digital. Fix: do not. A digital sensor needs none, and extra time only overexposes.
• Assuming equivalent exposures make identical pictures. Fix: they match in brightness only. Depth of field and motion change with every row of the table.
• Forgetting the aperture option. Fix: on long film exposures, one stop wider saves more than a stop of time, if you can spare the depth of field.

Try this: see the law hold, then see film break it

At dusk, put a camera or phone in a manual or night mode on a tripod and frame a dim scene. Find settings that give a good exposure of about 2 seconds, then make four frames that are all the same exposure: 2 seconds, 4 seconds one stop smaller in aperture (or half the ISO on a phone), 8 seconds two stops smaller, and 1 second one stop wider. It takes about ten minutes.

Judge the result by comparing the histograms. On digital they should line up within a small fraction of a stop, apart from any change in light as dusk deepens, while depth of field and any moving water or traffic look different. That is reciprocity holding. If you shoot film, repeat the set on one roll and add a fifth frame at the time the film’s data sheet gives for that 8-second reading (about 15 seconds with an exponent of 1.31). After developing, the 8-second metered frame will have the thinnest shadows, and the corrected frame will come much closer to the 1-second one: reciprocity failure you can see.

Want to test what you know about shooting film, from exposure to development? Try the film photography quiz.

Frequently asked questions about reciprocity

What is reciprocity in photography?

It is the principle that exposure equals light intensity multiplied by time. Because of it, you can trade a stop of aperture for a stop of shutter speed, or of ISO, and keep the same picture brightness.

At what shutter speed does reciprocity failure start?

For many films it becomes noticeable at metered times past about a second, but the point varies widely: some films need nothing until about two minutes. At the other end it appears only at extremely short durations, around 1/10,000 second and shorter.

How do I calculate reciprocity failure?

Use the data sheet for your film. If it gives a formula, raise the metered time in seconds to the published exponent: with 1.31, a metered 30 seconds becomes about 86. If it gives a table, read the corrected time beside your metered time and bracket long exposures.

Do digital cameras have reciprocity failure?

No. Digital sensors respond to light in a nearly linear way at any practical exposure time, so the metered time is the right time. Long digital exposures bring noise and hot pixels instead, which long exposure noise reduction or dark frames deal with.

Is it better to open the aperture or add time?

Opening the aperture saves more than it seems on film, because it shortens the time before the correction compounds: with an exponent of 1.31, one stop wider cuts the corrected time by about 2.5 times. Add time instead when you need the depth of field or the extra smoothing of moving water and clouds.

Does reciprocity failure make film overexposed or underexposed?

Underexposed. The film responds less than the meter predicts, so a frame given the metered time comes out thin, especially in the shadows, and with higher contrast than the scene.