Reviews
Working Within Exposure Limits in Digital Photography
by Dr. Glenn Rand, Brooks Institute of Photography
When we really think about photography, we get down to a singular point, that light is critical. Looking at the final image, when successful, we seldom see the light, yet without accurately capturing the light the image is likely to fail. It makes sense to set the first priority finding a way to relate the light’s energy to the final image. We can think of this as matching the light in the scene or on the controlled set to the capabilities of the camera and to the way we will present the final image. This is called measured photography.
Too often we believe that if we use a computer, either on our workstation or in the camera, we are controlling our process. Certainly since the modern digital camera has more computing power than the moon-lander we can think that this is the only piece of equipment we need to adequately capture the light in the scene or set. With the advanced digital camera, we can be lulled into believing that the camera can do it all. The modern digital camera has metering, white balancing, autofocus, autobracketing, etc. So what is missing?
What becomes clear is that control of the photographic process is what separates our efforts from others using mobile phone cameras. Control is what separates you from others who are “camera operators.” The issue is not if you can move away from accepting what the camera gives, but how to controll the process that sees the camera as it should be, a tool in your creative toolbox.
Since digital cameras rely on their sensors, we must deal with the limitations of these individual chips. The critical limit in digital capture is the inability of digital sensors to accept overexposure. Unlike negative film, a digital sensor has an absolute limit for maximum exposure. Once an area of the sensor has accepted its limit, no more exposure is possible. At this point either the energy from the light is ignored or it ‘blooms’ to adjacent pixels. Unlike negative film, we have no bright exposure latitude.
After the sensor has captured light, each system uses its own proprietary image array processor to interpret the image. Even with RAW files there is some image-processing happening in the camera. With JPEG files there is also interpolation and compression of the files before they are exported to the memory card. These operations in the camera can also affect exposure.
Some individuals say that you can correct the image’s deficiencies with post-capture software if you photograph in a RAW file format. While there are powerful tools within imaging software, they cannot put into an image detail that was not originally captured at the time of exposure. Neither can you reduce the intensity of a pixel that has been overexposed to allow lower value detail to be seen. Just as with transparency film, if you make a mistake and overexpose highlight detail it will be lost and look as though it is white.
Further, many other photographers believe that they can accurately control exposure by looking at the histogram produced on the camera’s LCD. But at best, the small size of the LCD and its lack of tonal selection or capture numeric measures leave this method suspect. The camera’s histogram is not accurate and only gives an indication of the fit of the light’s dynamic range within the limits of the sensor. When dynamic range surpasses abilities of the sensor to capture the energies there is a loss of image information. This is particularly true with bright light even if the shadow details can be lightened.
Depending on the demands for accuracy of output of the image, there are other concerns. Most important is your desire to have the output meet your expectations as a portrayal of the scene you photographed. If you are printing your image on an inkjet printer that can only reproduce a seven-stop range, then image detail beyond the printable range will be lost. This is known as “clipping.” This can also be considered as a point for the sensor. For this discussion we will use the term of “Clipping Point” as the cutoff point on a sensor, where the sensor has collected either not enough energy from the illumination of the scene to be recorded other than noise; or on the high end, that the energy was too great for the sensor to capture all the energy. In this latter sense the image may “bloom.”
In purely scientific use the clipping points define the dynamic range of a capture device. For this piece the idea of dynamic range will refer to the range of tones from the lowest useable detail area in the scene to the brightest useable detail limit. While many cameras produce files with 16-bit per channel (16 levels of intensity in each Red, Green and Blue channels in the image) we use 8-bit (0-255) to discuss the light intensity in any channel. In order to assure the usability of any portion of the image we interpret the maximum useable portion of the range as a digital value of 242 and the lowest value of 12. This allows for a moderate amount of noise in the darkest areas of the image and the ability to define detail in the brightest area of the image.
These three prime factors, sensor limits, light dynamic range and clipping points (capture limits), are needed to control maximize exposure. To achieve a high level of exposure control we need to address our exposure within our digital imaging world defined in these three areas. The most efficient and expeditious way is to use light metering technology. A light meter can do more than the in-camera metering system. Because a meter can be designed to read both incident and reflective light it can relate to the total light space that we are photographing.
Beyond the basic metering, advanced meter that will enhance your exposure control requires these three specific functions… a method to coordinate these aspects with cameras and light sources producing your digital photographs; determination of the dynamic attributes of the light; and a way to relate capture of light that you will use to fulfill your photography. While in this piece we will only deal with exposure profiling, output needs to be considered for the best final results.
To build a system approach to better exposure with digital cameras using a light meter we need to construct an intermediary function that will allow us to work within a changing light environment. The term we use in computer management is a “profile.” So we need to construct an exposure profile that becomes the interface between the light environment as measured with the meter, the camera that will record the light reflecting from the subject, the computer that will process the image and the output device that will fulfill the photograph.
The easiest part of profile building is the light and this is a good starting point. Different light sources will have varying spectral distributions or color biases. These can have vastly differences on the effective exposure of the digital sensor. Just as daylight and tungsten films were designed for different light sources, the sensor even with white balance has slightly different exposure capabilities when photographing under different light sources. You will need a different profile for each light source you use to maximize exposure. The light energy also is distributed across an intensity range, the dynamic range, and this needs to be part of any exposure profile.
If you use more than one camera, you will need to have a profile for each camera. Even though Fuji and Kodak manufacture daylight films, the films vary in their sensitivities and color biases and expose differently. In this same way there are different exposure capabilities between CCD and CMOS chips. Beyond the difference created by different sensor types, each camera system has its own proprietary array processor that handles the light gathered on the sensor.
Some chips are faster than others and most have variable ISO settings. The variability is controlled in several ways but all methods of adjusting ISO affect exposure beyond the obvious speed setting. That means that if you are using your digital camera at different ISO settings another factor has been added to the profile for maximizing exposure.
Within professional systems you might also choose to save the captured images in JPEG or RAW file formats. This is more than an issue of compression because JPEG interpolate color while RAW does not. There are advantages to each and each choice of file format to be used adds another difference to how exposure is handled and maximized through an exposure profile. Even within any format there is an issue of what bit depth is captured. If a sensor is capable of capturing 14 stops of dynamic range in 8-bit mode it will compress exposure to fit and in 16-bit mode it will expand the file.
The idea of profiling is to facilitate communication between the parts of the photographic process. Key in developing this exposure profile is the tool in the middle, the light metering. We need to build a series of look-up tables based on light values. These look-up tables are coordinated measurements from the camera and printers used with various light sources.
Because of the many parts and variability of the exposure profiling process the testing and profile building can be complicated. However, with the Sekonic DigitalMaster L-758DR we have a tool that allows for the construction of these profiles in its memory. The meter allows the creation and storage of the profiles and communication between portions of your photography to maximize exposure.
In this article we will not go through the exact steps of creating and using exposure profiles, we will give a very brief overview of the process. Creating an exposure profile can be broken down into; 1.) establishing the camera’s baseline, 2.) reading outcomes of camera testing to establish camera’s exposure performance, 3.) adjusting the camera’s exposure information to achieve desired capture, 4.) storing the exposure profile and 5.) coordinating the information with camera limits.
At any ISO, the camera needs to be custom white balanced for the light being tested. To baseline the camera we use a known target that has a black, white and 1/6th stop stepped gray scale. This allows us to expose and find the true exposure index of the sensor as well as establish dynamic range. To effectively find the true exposure index we bracket exposures over a 7-stop range (+3 to -3 stops). Once the files have been moved into Adobe Photoshop® they can be opened and using the Info Window the level of the gray scale that most closely meets the expectations of correct exposure is what we are determining. This will be the step that is closest to 118 as a green (G) level. The metadata from the image is then used to establish the exposure index at the set ISO. The correct ISO for the sensor is determined from how much variation there is from the 118 value being found on the ISO exposure file on the 18% gray tone step. For each step variation from the 18% step is 1/6th ISO stop from the camera setting. For example if the 118 level is two steps brighter than the 18% step and the ISO was 100, the effective exposure index is ISO 80. In this example the exposure compensation is .33. Since variations in ISO change the exposure profile at least two ISO’s must be tested to use the range of ISO’s available on your camera.
Once the exposure index is established, the dynamic range can be measured by reading the white, black or gray areas of the target. These may be on various files, since your corrected exposure index file may not have the total range captured by the sensor in your camera. The variation of the light from the patches of your target needs to be measured at the time of exposure and their stop variation noted. For example, if the stop variation is 7-stops when exposed and the sensor gathers more than 7-stops of dynamic range then the information on dynamic range will come from more than one file. The key numbers in the Info Window are slightly less than maximum for the white (239-242) and slightly greater than minimum for the black (4-12). If white falls at 242 in the +2-stop exposure and the black is 10 on the –2-stop file and you had 7-stops between the white and black in the original metering then your camera captures a 12–stop dynamic range (7+ 5). (+2, +1, 0, -1, -2 = 5 stops)
This information provides the basis of the exposure profile. The Sekonic DigitalMaster L-758DR using proprietary software can acquire and develop the profile information from the computer via USB for storage and future use. The meter allows multiple cameras to be profiled and saved. With the exposure profile on board the meter can then allow you to get accurate exposure information based on the performance of your camera while allowing you to see where the white limit will be reached and if any clipping will occur.
To use the developed profile there are three controls that can be applied to exposure. First, with the exposure index established the light requirement for a middle-gray or incident meter reading gives a center point value for correct exposure. Just as any starting exposure value, it can be replaced with an equivalent exposure where needed. Also, since the profile are based on the type of light source the exposure calculation takes into account the biases of the light. Second, reflective meter readings are taken of the brightest area of the image that will be white and the area that will represent black. These readings establish the dynamic range of the light and when compared to the middle-gray reading will show if the white level is beyond the ability to capturing the bright detail without exceeding the capability of the sensor to record the intensity. If the intensity of the white is beyond the sensor’s limit for recording then an exposure change can be used that will bring the white into capabilities of the sensor.
With your exposure system profiled you have options to maintain accurate exposure while freeing you up to be creative without worrying about the outcome of your exposure.