The Online Photographer has posted an excellent article, More on Dynamic Range, on the range of brightness in a scene, how it is captured as a photographic image, how to fit that range into the range of lightness levels recorded by the camera and express that range in the rendered medium, whether a JPEG image viewed on a monitor or a paper print.
He's right that dynamic range is the most abused, misused and poorly understood term in digital photography. It's the only short hand we have for "range of brightness values" or "range of tonal values," which are both going to give your fingers cramps if you write them often enough.
There is a lack of understanding by many photographers about the basic process of recording an image and producing a visible print from it. There are crucial, but precise, distinctions to be made, which took a long time and much expertise to establish in analog photography, so the confusion is not surprising.
The first thing to consider is the range of brightness in the scene (which the Online Photographer article demonstrates and discusses). It may seem obvious to some, but is often counter-intuitive, that a distinction exists between the range of measurable brightness values in the scene (and remember, most are reflected light, but some is direct from light sources or specular reflections, for purposes of exposure, it is good to consider only reflected light and not specular highlights, since they do not contain any detail or information) and the representation of those values as tonal values in the recording medium (in the camera, film or sensor).
The difference between the range of brightness in the scene and the translation of the brightness levels into tonal levels recorded by the camera (density in film, tonal levels in digital) is not immediately obvious, but the camera does not record brightness, but some analog of it, clumps of grain or numbers. To see the picture, the range of tonal levels must be translated back into a range of brightness values. We do this when printing a negative to photographic paper or viewing a transparency film slide through transmitted light (a projector or lightbox).
In the digital realm, the image rendered from raw capture data or printed to paper is the output, which must be translated into reflected or transmitted light so we can view the image. A complication in digital photography is the JPEG image, which places limitations on the original data. It would not matter either, if it were a TIFF image, since all images rendered from capture data have contrast curves applied to fit the image within the range of tonal values the format is capable of storing and to be "pleasing" to the eye. Linear-data is not pleasing to the eye because it contains too large a range of tonal levels and corresponding brightness range. It won't "look" like the original scene as the eye saw it.
When you are talking about dynamic range, you first need to ask, which range? Is it the range of brightness levels in the scene, capable of being captured by the sensor as input, capable of being rendered to output? Is it the range of brightness or tonal values you are considering?
The scene has a range of brightness values.
The recording medium (film or sensor) has a range of brightness values it is senstive (ISO comes in here) to and a range of tonal values it uses to express those values. The brightness levels are translated to those tonal levels (whether represented by density in analog film or by numbers in digital data).
The output medium has a range of tonal levels it is capable of storing and expressing as brightness values when viewed.
The complications come because of the need to match the range and step of tonal levels in the input to the output. Further complicating things is that the JPEG image has its own set of curves and translations, when printed the printer paper and inks place their own set of limitations and curves on the translation. The environment in which the image is viewed has its own limitations and effects on the brightness values percieved.
The capability of a camera or sensor cannot be judged by looking at a random example of output from a camera's JPEG engine. That would be the equivalent of judging a film by the quality of processing and printing from a randomly selected corner drugstore.
Comments