High-contrast scenes create competing demands on pupil size, producing local over- or under-exposure

What it is

When a scene contains both very bright and very dark regions at once — strong backlighting, a bright warning badge embedded in a dark-themed night interface, high-dynamic-range content with a huge brightness spread — the pupil can only settle on a single aperture; it has no way to adjust separately for different regions of the scene. Constricting to suit the bright areas leaves dark-area detail under-exposed and hard to make out; dilating to suit the dark areas makes the bright areas glaring, washing out their detail. This is not a matter of the display's dynamic range being insufficient — it is a direct consequence of the pupil being a single aperture that must strike one compromise for the whole visual field.

Why it happens

The signal driving the pupillary light reflex is an integration of photoreceptor input from across the visual field — roughly a combined luminance from the central region and a surrounding area — rather than an independent response computed point by point. In other words, pupil aperture is fundamentally a global compromise reaction to the current field's overall luminance distribution; it cannot process different regions separately the way a camera's local exposure or multi-frame HDR compositing can. When a scene's brightness range exceeds what a single compromise aperture can accommodate, no aperture choice is optimal for every local region: too large and the bright areas look over-exposed, too small and dark-area detail becomes hard to perceive — the two ends trade off against each other and cannot both be satisfied at once.

Studying it

A common approach has participants view scenes with different bright/dark contrast ratios, measuring their stable pupil aperture while also measuring subjective discriminability or recognition accuracy for both bright-area and dark-area detail, to test whether the pupil's compromise aperture trades off against detail discriminability at both ends. Another approach systematically varies the area proportion occupied by bright versus dark regions in the field, examining how that area weighting shifts the eventual compromise aperture. Common independent variables are the luminance difference and area proportion between bright and dark regions; common dependent variables are the stabilized pupil aperture and recognition accuracy for bright-area/dark-area detail.

Where it stops holding

This limitation only becomes evident when a scene's brightness contrast exceeds what the retina can process simultaneously across its dynamic range; ordinary interfaces under uniform indoor lighting rarely trigger this conflict. In addition, peripheral vision does not carry the same weight as the central fixation area in driving overall pupil aperture, so the specific composition of the visual field (whether the bright region falls near the point of fixation) shifts the eventual compromise aperture — it cannot simply be predicted from the scene's overall average luminance.

Applying it

• When designing high-contrast scenarios (bright warnings embedded in a dark theme, night-driving displays, access-control or kiosk screens used in strong backlight), avoid putting the most critical information exclusively in an extremely bright or extremely dark region; keep key content in a mid-luminance area whose brightness does not diverge too sharply from its surroundings, reducing reliance on the pupil's single compromise aperture.
• Where high-contrast content genuinely cannot be avoided, consider presenting the most important information from only one end (bright or dark) at any given moment, rather than requiring the user to read critical content from both ends at the same time.
• Verification: under the target lighting conditions, measure users' actual recognition accuracy for critical content in both the dark and bright areas of the scene; if both ends show a noticeable drop, the current contrast range exceeds what the human pupil can accommodate simultaneously, and the brightness spread should be compressed or the content separated in time.

Related

• Same group: A1.20.1 Pupil diameter automatically contracts and dilates with ambient light to regulate incoming light · A1.20.2 Pupil adjustment is slower than the brightness jumps electronic displays can produce · A1.20.3 Pupil size also sets depth of field, with a smaller pupil raising tolerance for focus error · A1.20.4 Pupil response is also driven by cognitive load and emotional arousal, not light alone
• Nearby: A1.04 Contrast sensitivity
• Site search: pupil adaptation conflict · high dynamic range viewing · local overexposure · local underexposure