At the time of writing there is no standard method for characterizing digital cameras, but useful results can be obtained by the adaptation of sensitometric methods developed for the examination of silver-based photographic systems.
The light-sensitive element of a digital camera, usually a charge coupled device (CCD), unlike a photographic film, is not something that can be taken away and examined. Any investigation of its sensito-metric properties has to be carried out in situ, within the camera. The elementary methods discussed above have to be used but if acceptable accuracy is to be achieved it is usually necessary to have calibrated grey scales or sensitometric wedges. Alternatively the need for a wedge can be avoided if the shutter of the camera can be controlled and, preferably, the lens removed. The camera can then be set up so that a suitably controlled light source is allowed to illuminate the shutter and a series of exposures can then be made over as wide a range of exposure times as possible.
Calibrated reflection grey scales are available commercially at reasonable cost and should have as large a density range as possible, certainly at least 2.0. A transmission neutral wedge of density range 3.0 would be better in determining the dynamic range, or useful exposure range of the camera. The grey scale should be evenly illuminated and photographed using the digital camera under study. The evenness of illumination can be tested by superimposing a sheet of white card and repeating the exposure. The camera will then act as a photometer and characterize any unevenness due to illumination of the grey scale, or within the camera itself.
The output of the CCD is customarily stored within the camera and can be accessed by a suitable computer equipped with appropriate software, sometimes, but not always, including a driver for the camera. What is required is the output in pixel values from 0 (darkest black) to 255 (lightest white) for each step of the grey scale. A graph of pixel value against relative luminance of grey scale is shown in Fig.15.45. Because most calibrated grey scales possess fairly constant tonal increments, the steps are normally calibrated in reflection density. In order to calculate the relative luminance scale we use the equation shown earlier in the chapter which defines density, but modified for reflection density
ylr and rearrange it to calculate Ir/Ii as a measure of relative luminance where values of 10DR may be conveniently looked up in Antilogarithm tables or directly evaluated using a hand calculator.
The transfer function obtained illustrates the response of the CCD camera to exposure level and has been fitted in the diagram by a convenient mathematical function, giving a smooth curve. It shows a reduction of gradient as the limiting white reproduction is achieved, but it is difficult to interpret in the light of existing knowledge and habits of thought about photographic characteristic curves. If, on the other hand, we plot log digital value against exposing wedge density (a logarithmic measure) we obtain, from the same experiment, the result shown in Figure 15.46.
This is a transfer function resembling a low contrast reversal photographic characteristic curve and can be interpreted in a similar way. First, the data are more evenly distributed than in the linear scales of Figure 15.45, and closely fit the linear regression drawn through them. The equation of the straight line appears on the graph. The linear response effectively runs up to the maximum possible, at 255, from a minimum at 30. This minimum persisted at higher
Figure 15.46 The characteristic curve of a CCD camera plotted as log digital value against exposing wedge density
Figure 15.46 The characteristic curve of a CCD camera plotted as log digital value against exposing wedge density wedge densities, representing the minimum output possible with the camera in its normal configuration, with lens in place and essentially photographing a suitably illuminated sensitometric step wedge. Some of the minimum digital value probably originates as flare and some as detector noise. The former can be largely avoided by removing the lens and using time-scale sensitometry. The latter will remain under these circumstances. The linear response in Figure 15.46 gives a useful exposure scale of approximately 2.2 log units, equivalent to 7 stops.
The mean gradient of -0.57 evaluated from the straight line shown in Figure 15.46 is approximately equivalent to that of photographic negative films, when conventionally developed, but takes no account of contrast modifications which arise in later stages of the digital imaging chain. The output device for example, commonly a computer monitor, will increase the effective contrast considerably, and visual inspection usually reveals sufficient contrast in the final reproduction under normal viewing conditions.
This study of a digital camera shows the power of quite simple sensitometry. The test subject was a transmission step wedge. The camera output was evaluated with computer software commonly used in digital imaging.
In many applications the nominal calibration of the neutral step wedge is adequate. If greater accuracy is required then it is necessary to calibrate the grey scale separately using equipment not commonly available outside a laboratory, and to determine the spectral properties of the illuminants used. This also requires expensive laboratory equipment. In fact the full resources of photographic sensitometry are required for the precise characterization of both silver-based and digital photographic systems.
Was this article helpful?
Insider Secrets Revealed By the Pro Showing You How to Become a Professional Photographer! Discover The Secret Tips & Techniques On How To Be A Professional Photographer, Start Producing High Quality Pictures and Skyrocket Your Photography Business Income Revenue To The Roof TODAY! You're About to Discover the Powerful Strategies and Method to Start Taking Sharp, Clear and High Quality Pictures Like the Professional Photographer Without Paying a Single Penny to the Expert!