dynamic range

Measuring the range

Dynamic range describes the interval between the two limits of a recording medium: the highlight level at which it can hold no more light — saturation, or clipping — and the shadow level below which detail disappears into random electronic or chemical noise, the noise floor. The distance between these bounds is customarily counted in stops, where one stop is a doubling or halving of light, so a range of ten stops spans a highlight roughly a thousand times brighter than the faintest recorded shadow. Because the measure is a ratio rather than an absolute brightness, it is independent of how much total light a scene contains and depends instead on how faithfully the medium separates the brightest from the darkest tones it can capture at once.

Film and digital compared

Photographic film and the digital image sensor reach their limits by different physics. A film emulsion responds to increasing exposure along a characteristic curve that bends gradually toward its maximum density, so overexposed highlights compress and roll off rather than terminating abruptly; this soft shoulder gives negative film a wide latitude for excess light. A sensor’s photosites, by contrast, fill with charge linearly and then clip sharply once a photosite reaches full capacity, producing a hard, featureless white. At the dark end the sensor’s floor is set by read noise and thermal noise, while film’s is set by the base fog of the unexposed stock. Modern cinema sensors are commonly rated in the range of roughly thirteen to fifteen stops, comparable to or exceeding motion-picture negative, though the two distribute their latitude differently across the tonal scale.

Exposure, highlights, and shadows

Dynamic range constrains exposure by fixing how much of a scene can be rendered with detail at a single setting. When a scene’s own range of brightness exceeds the medium’s, the exposure decision becomes a choice about what to sacrifice: exposing to protect highlights pushes shadows toward the noise floor, while exposing for the shadows risks clipping the highlights beyond recovery. Because a sensor’s response is linear, most of its recorded levels fall in the brightest stops, so exposing as brightly as clipping allows — placing the signal well above the noise floor — retains the most shadow information, a practice known as exposing to the right.

Log encoding and raw capture

To carry a wide captured range through recording and post-production, digital cinema cameras rarely store the sensor’s linear values directly. Log encoding applies a logarithmic transfer function that allocates roughly equal code values to each photographic stop, mirroring the perceptual spacing of film density and letting a wide range survive within a limited bit depth without visible banding. Raw capture goes further by recording the photosites’ original values before demosaicing or a fixed tone curve is applied, deferring those decisions to post-production and preserving the full range the sensor recorded. Both approaches separate the range a medium can capture from the narrower range a display can show, so that highlight and shadow information survives for later grading. The count of distinct levels within that range is a matter of quantization and bit depth, and so relates to but is distinct from spatial resolution.