High-dynamic-range imaging
HDR imaging will change the way you make exposures. HDR imaging will change the way you process your exposures. HDR imaging is the future of photography—to a large extent, it’s here now, but it will advance rapidly. At a minimum, you need to know what HDR imaging is and get ready to practice it in the near future. You may want to start practicing it now.
HDR imaging represents a radical departure from traditional forms of photography. It cuts straight to the heart of one of the most fundamental challenges of making photographs, fitting the dynamic range of what we see into the dynamic range of a capture device first and an output device second. HDR offers photography where all luminance values in a scene are recorded and exposure is set after capture. HDR imaging is so fundamentally different, that core concepts in photography need to be reconsidered.
Dynamic Range
What is dynamic range? In imaging, dynamic range (DR) is the highest overall level of contrast found in an image. In other fields, such as in the audio industry, dynamic range is used to describe similar phenomena. In audio, DR is defined as the logarithmic ratio between the largest readable signal and background noise. Imaging DR is akin to audio signal-to-noise ratio.
Consider an image a signal. Every image also has some noise. The values used to specify dynamic range can be charted on multiple scales. Whatever language is used to describe this phenomenon, two critical factors must be addressed—the total range of brightness and the fineness of the steps used within the scale.
 |
 |
 |
| 1) Dynamic range in EV and CR compared 2) Curve tone responses Red: CCD RAW Green: Film Blue: Gamma encoding Black: Processed RAW result after gamma 3) RAW, HDR, LDR capture-merge histogram comparisons 4) Final image |
The EV scale makes it easy to compare the ratios rather than the big numbers of logarithmic progressions; each successive EV rating represents a doubling of values. The exposure value (EV) scale has been used by photographers for ages. The International Organization for Standards (ISO) defines EV 0 as an aperture size of ƒ/1 and a one-second exposure time. The same EV can be achieved with any other combination of ƒ-stop and shutter speed that produces the same amount of light.
The contrast-ratio scale specifically delineates values; when you use this rating, you instantly see how much greater each step in a progression is than the previous one because the numbers are so much bigger. You can convert EV to contrast ratios or vice versa with the right formulas. 2 (power of EV) = contrast ratio (2*8=256 for a contrast ratio of 256:1) or EV=log10 (contrast ratio) *3.32 (log10) (4000)*3.32=12EV.
Dynamic range, gamut and bit-depth are often confused. Though related, they’re all different. Dynamic range refers to a total range of luminosity values. Gamut refers to a total color capacity, including saturation. Bit depth refers to the number of points of data between values or the fineness of the increments in the scale. Just because an image is wide-gamut doesn’t mean it’s HDR or has high bit depth, but it will contain more and potentially better data if it does. Likewise, just because an image is HDR doesn’t mean it’s wide-gamut and has high bit depth, but it will contain more and potentially better data if it does. You can’t convert low-dynamic-range, small-gamut, low-bit-depth information to high-bit-depth, wide-gamut, high-dynamic-range information. To get it and use it, you have to capture high-quality information upon exposure and preserve it throughout your workflow.
How The Camera Sees
Like the human eye, film has a nonlinear response to light. For film, we adjust the EV to fit the amount and contrast ratio of the available light into the most useful area of its curve response. Using film, you expose generally, and when compromises need to be made, you favor shadows or highlights. Details lost at the point of capture are irrecoverable.
Unlike the human eye, CCDs have a linear response to light. They simply count photons, with no scaling. Consequently, in linear capture (RAW), half the data in the file is contained in the top EVs of the tonal scale, and the quality of the data in the lowest EVs is comparatively poor (susceptible to noise and banding). RAW files without conversion look very dark. When converted, a tone curve is applied (gamma encoded) to make them appear normal. The images are mapped to an output-referring standard.
Camera capture (film and CCD) offer roughly 8 EV compared to the eye’s 14 EV.
Read More...