How to determine the dynamic range of your camera


Dynamic range refers to the difference between the darkest and brightest areas of the subject of a photograph. In nature, there may be more than a thousand-fold difference in brightness in a high contrast subject. How much of this can your camera capture? If you look at the histogram on the LCD screen on your camera it probably only shows the 5 brightest stops of dynamic range (5 stops equals 2^5 or 32-fold difference). Since the middle stop shown is the midtone, it is true that you can only capture 2.5 stops above this (see below), but how many stops can you actually capture below the midtone?

One way to find out is to look it up. You can go to the website and look up your camera. For example, my camera is a Canon EOS 7D. According to dxomark it has a dynamic range of 11.7 stops, which represents about a 3000-fold difference between the darkest and lightest areas. This means I should be able to capture 9 stops of tonality below the midtone. But can I, and if so what is the quality of the capture?

When you consider how a camera sensor works, there is clearly an upper limit to image capture. A midtone is defined as 18% maximum brightness for the sensor which is 2.5 stops below the maximum. Above the maximum, the sensor is saturated and cannot register further increases in brightness. Below the midtone there is no theoretical limit to what can be captured as long as enough photons strike the sensor to cause charge to accumulate. However, there is a practical limit when photons from the image become so few as to become indistinguishable from the random accumulation of charge which occurs on all sensors. At this point no distinguishable image could be discerned. As one approaches this point, however, image quality would be expected to deteriorate.

You can test this on your own camera, as I did on mine. I chose a subject, in this case an oak tree trunk, which had an approximate midtone brightness on an overcast day. This figure shows the black and white image of the trunk with an 18% neutral grey card. As you can see, the trunk is slightly darker than neutral grey.



I took a spot meter reading off the trunk which assigns a midtone value to the trunk and took an exposure. Then I took exposures 1, 2, 2.5, and 3 stops overexposed. I also took exposures 1-5 stops underexposed. I loaded them in Adobe Lightroom and used the exposure adjustment to bring each image back to midtone. For example, if I underexposed the image 3 stops, I increased the exposure 3 stops in Lightroom. 5 stops is the most you can adjust exposure in Lightroom which is why I didn’t take further underexposures. I also looked at the overexposure warning on the LCD screen. The 2.5 stop overexposed image showed scattered highlight warnings, and the 3 stop overexposed image showed extensive overexposure as expected. The figure below shows reduced-size versions of the adjusted images. + or – indicates how many stops the image was over or underexposed from the midtone reading.

Focusing on the tree trunk, there is at least a reasonable image capture over the total 9 stops of range I tested. Looking carefully, the images underexposed by 4 and 5 stops may look a little ragged. There is also a slight reddish color cast which I believe is due to the sensor being more sensitive to red light at low exposures. Looking at the overexposed images, the tree trunk looks reasonable even at 3 stops overexposed. However, if you look at the green leaves in the images overexposed 2.5 and 3 stops, you see a phenomenon referred to as channel-clipping. This occurs because in the properly exposed image, the green color channel in the leaves is brighter than the red or blue channels as expected in a green leaf. When the image is over-exposed, the green channel is “clipped” meaning that as exposure time increases, no more green data can be recorded. However, the red channel, being less intense to start with, is not yet clipped, so increasing exposure increases the amount of red channel information relative to green, causing the leaf to turn a yellow-orange color. You still get a picture of the leaf; it is just the wrong color.

The next figure shows a small area of the tree bark at 100% size.

It is apparent that the images underexposed by 4 and 5 stops are “grainy” due to higher amount of sensor noise relative to the image. There is slight grain in the 3-stop underexposure.  When you increase the exposure in Lightroom you amplify both the real image as well as the noise. As an aside, this is also what happens when you increase the ISO setting on your camera. When you increase the ISO by two stops (ISO 400) for example, you are really underexposing by 2 stops and then increasing the exposure brightness via the in-camera software. This is why high ISO exposures often look grainy and require more noise reduction in processing.

The 100% images of the 2.5 and 3-stop overexposed images show localized flat grey areas which represent highlight clipping, where the sensor maximum had been reached and differences in tonality could not be recorded. As you can see, adjusting the exposure in Lightroom doesn’t fix the problem.

My Canon EOS 7D has a nominal dynamic range of 11.7 stops. Practically, the best exposures are between 2.5 stops above and 3 stops below the midtone stop for a total of 6.5 stops. This does not mean that I am limited to taking pictures within a 6.5 stop range. Some graininess may be tolerated in deep shadow areas with limited detail. Noise-reduction software can be used to reduce the grainy appearance in lower detail shadow areas. Noise is less apparent in smaller prints and LCD screen images, and might only be a problem in large prints. Exposing the image to the right (see How to Take a Proper Exposure) reduces the likelihood of image degradation in darker areas of the image. However, in very high contrast scenes, consideration should be given to taking multiple exposures and combining with an HDR program to assure good capture of detail in shadow areas.

Back to Photography Techniques