Reciprocity is a property of physics describing potential errors in flow rates. It applies to liquids, hydraulics, radio waves, sound waves, electrical magnetic flux, and yes, light. Thus it applies to image recording whether it is film or digital. Reciprocity errors typically occur at the extreme high or low flow rates near the design limits of a system.
One way to describe this is to use the flow of water from a source to a measuring cup. In this case the source will be a simple faucet. In the illustration above, a normal flow rate will fill the measuring cup accurately.
If the water pressure is too high, the handle is opened too far, or the opening is too large, some water is lost. It will splatter out of the faucet, bounce off the sides of the container or even overflow the water already in the container. It may be very difficult to turn off the faucet before the container fills, so some will spill out.
If the water pressure is too low, the container fills very slowly. Some water may be lost to evaporation. The walls of the container might absorb some of the water. If the ambient temperature is high the rate of evaporation can even exceed the rate of fill. If the humidity is high and the temperature is low, condensation might add more water than the faucet.
The summary is easy. At normal flow rates within the design limits of the faucet selected and the container used for the measurement, the measurement is accurate and linear. Near the design limits of either the faucet or the container, the measurements will be less accurate and non-linear.
These same considerations apply to the digital sensors in your camera.
The water is now the light, the measuring cup is now the sensor, the aperture is now the faucet, and the shutter is now the handle. All of the aforementioned principles still apply.
Photons of light are converted in a silicon well to electrons so they can be measured. The walls of the etched silicon container are more porous than the Pyrex glass. Like the measuring cup, the silicon well is a finite and unvarying size, though much smaller. Thus the design limitations are more stringent and the manufacturing tolerances are more critical.
When the well is full, it will overflow. When too many photons arrive too fast, they will bounce around in all sorts of unwanted places. When there are too few, stray electronic noise and photons of stray light will overwhelm them. Some that were collected may simply be absorbed by impurities in the materials. All of this is subject to variations in the temperature and other environmental conditions. With longer shutter durations or rapid sequence shooting, the camera components themselves will generate heat.
We do not need to dissect all the minute details of automotive construction, but we do need to keep it in the middle of the road. Same here. Everything is designed around middle tones of light and time-honored photographic exposure settings. If we try to base our exposures on either the brightest highlights or the deep shadows we are approaching the design limits and will experience reciprocity.
We will shift the exposure settings to accommodate unique conditions of the scene. Nothing more or less than what Ansel Adams teaches with his Zone system. We should check the digital histogram to validate our assumptions. But the histogram is not a micrometer. We should resist the temptation to micromanage the highlights or shadows with it.
All cameras are not created equally. Digital or film, you get what you pay for. Some digital cameras have very small sensors, some are twice or three times as large. Some have excellent dynamic range. Some donít. Some do better than average in very low light or high ISOs. Some donít. They have different image plane sizes, optics, mechanical components, and features. Some will capture extreme highlights somewhat faithfully. Some wonít. Some will handle long exposure times quite well. Most wonít. Take some time to experiment and learn the limits of your equipment. Make sure you understand basic exposure techniques. Then the histogram will provide useful validation in the field.
That is just my two cents. I hope you also gained some new insight from this article. If you have any comments, or suggestions, I would welcome your input. Please send me an Email.
Rags Int., Inc.
204 Trailwood Drive
Euless, TX 76039
September 14, 2004
This page last updated on: Thursday April 12 2007
You are visitor number 8,360 since 01/25/06