What is dynamic range, and why can't highlights and shadows both be preserved?
Camera dynamic range is the brightness range in one exposure, from unsaturated highlights to usable shadow detail. Learn about RAW latitude, highlight recovery, shadow noise, HDR, and how to read test results.
Camera dynamic range is the brightness range in one exposure, from unsaturated highlights to usable shadow detail. Learn about RAW latitude, highlight recovery, shadow noise, HDR, and how to read test results.
Dynamic range is a camera's ability to record usable detail in both the brightest and darkest parts of a single exposure. Once highlights reach sensor saturation, they usually cannot be recovered; shadows can be brightened, but noise will become visible. RAW files preserve more room for adjustment, but they cannot create light or detail that was never recorded.
How is dynamic range defined?
Photography usually expresses dynamic range in “stops”: the ratio between the brightest unsaturated signal and a dark-area noise threshold considered acceptable. Each additional stop doubles the brightness range. Imatest's explanation of dynamic-range measurement reports the range at different signal-to-noise ratio thresholds, which shows that “acceptable” is not a single universal standard. Engineering dynamic range, the range usable in a print, and the signal-to-noise thresholds used by review sites may all differ, so figures from different sources cannot be compared without considering their methods.
Dynamic range is also not the same as image contrast. A high-dynamic-range file can be mapped to a low-contrast image, while a low-dynamic-range scene can be given punchy contrast with a curve; recording capability and final presentation are two separate stages.
What is the difference between clipped highlights and shadow noise?
Once a pixel reaches its full-well capacity, increasing exposure records no additional tonal levels. Important highlights become difficult to recover once all color channels are saturated at the same time. What is called RAW highlight recovery may sometimes mean that one channel has not yet saturated and the software reconstructs information using color relationships; a pure-white area contains no real texture to retrieve.
Shadows, by contrast, usually still contain a weak signal, but read noise and photon noise make up a larger proportion of it. Brightening the shadows amplifies the signal and the noise together. For an explanation of where it comes from, see how image noise is produced.
Why do RAW and JPEG have different latitude?
RAW files preserve higher-bit-depth sensor data with less baked-in processing, making it easier to readjust white balance, curves, and noise reduction. JPEG files have already gone through tone mapping, sharpening, noise reduction, and 8-bit encoding, so extreme adjustments are more likely to produce banding or compression artifacts. For a detailed comparison, see the differences among RAW, JPEG, and HEIF.
But RAW is not an unlimited safety net. A saturated sensor, a moving subject, or an incorrect focus point cannot be fixed by changing file format. The in-camera histogram usually comes from the JPEG preview, so the picture style and white balance affect where warnings appear. In important situations, leave a sensible margin.
How does ISO affect dynamic range?
On many cameras, the sensor has greater highlight capacity near base ISO. After analog gain is increased, the same output ceiling corresponds to less charge in the sensor, so highlight headroom usually decreases. Some dual-conversion-gain sensors switch readout modes at a particular ISO, which can improve shadow noise; as a result, the curve is not always smooth.
“Expose to the right” should mean increasing actual exposure without saturating important highlights or compromising the intended shutter speed and aperture—not blindly raising ISO or forcing the histogram to the right. For exposure control, see how to use exposure compensation.
When should you use bracketing and HDR?
For static architecture, interiors with windows, and sunsets, you can shoot several exposures and combine their highlights and shadows. The procedure is to lock the composition, white balance, and aperture, then vary shutter speed for the bracket. Make sure the darkest frame protects the highlights and the brightest frame produces clean shadows. People, leaves, waves, and vehicles can create ghosting, so shorten the time between frames or accept the trade-offs of a single exposure.
The final HDR image still has to be mapped to the brightness range of a display or sheet of paper. A natural result depends on local contrast and color management; HDR does not mean compressing all shadows and highlights until they have the same brightness.
How should you read test results correctly?
- Confirm whether the test measures RAW or JPEG, and the sensor or the complete camera output; Photons to Photos' PDR charts, for example, publish the definition and measurement conventions for their Photographic Dynamic Range charts.
- Check the signal-to-noise threshold, normalized output size, and bit depth.
- Compare cameras using the same ISO definition and similar resolution, and watch for dual-gain transitions.
- Do not equate a difference of a fraction of a stop directly with a visible difference in image quality.
- Consider the lens, exposure, and real shooting conditions instead of ranking cameras by a single peak number.