Raising refresh rate lowers available brightness

What it is

Lift refresh from 72 Hz to 120 Hz and each frame shrinks. If, to avoid smear, the panel still only flashes a short pulse at the end of the frame, on-time is cut with the frame and integrated luminance (nits) drops. That is the refresh–brightness tradeoff: more temporal samples, fewer photons. HDR, low persistence and high refresh grab at the same peak current; they cannot all be maxed.

The topic is the photon budget, not whether flicker fuses, and not the retinal geometry of smear.

Why it happens

Perceived brightness is photons integrated over a window. Under low persistence that window is about the pulse width. Raise refresh, shorten the period; hold duty cycle (pulse / period) constant and the pulse must shorten too, so integrated nits fall in proportion. Hold the pulse’s absolute width instead and raise duty cycle, and smear returns to long persistence — brightness bought back the ribbon just killed.

OLED instantaneous current is capped; raising peak to recover brightness ages the panel and trips luminance limiters. LCD backlights can be driven harder, at the cost of bleed, heat and battery. Outdoors or on optical see-through, virtual luminance has a higher floor than a cinema, so a refresh bump greys the content first.

A spec sheet “120 Hz” that does not say at what peak nits it was measured is usually a dark-lab number. In a bright room the driver will often drop refresh to keep brightness; what the wearer sees is not the spec, it is the trade made for them.

Studying it

Spot-meter or a calibrated camera, sweep refresh at fixed drive current, once with pulse width held and once with duty cycle held, and plot nits against hertz. Then do a readability or contrast match under given ambient light.

Independent variables: refresh rate, pulse width versus duty cycle, peak-current cap, ambient light. Dependent variables: integrated luminance, contrast, “bright enough” ratings, whether the driver auto-dropped refresh.

Log whether auto-brightness / auto-refresh engaged, or the curve is the policy’s, not the panel’s.

Where it stops holding

Very dark night content has luminance budget to spare; refresh can go first. Optical see-through, outdoors, white UI: the trade shows immediately. A dim monochrome indicator barely feels it. Maximum nits in a black lab are not usable nits in an office. A video-see-through camera has its own exposure; if raising display refresh forces the virtual overlay darker, virtual–real match breaks first and the complaint is filed as “AR looks fake”, not “refresh is too high”.

Applying it

• Write refresh and brightness as a visible pair: warn that peak luminance falls when refresh rises, or default to a lower refresh in bright rooms.
• Do not advertise 120 Hz and cinema-grade HDR peak without conditions; list “refresh sustainable at xx nits”.
• If an automatic policy drops refresh to hold brightness, show which step is active, so “settings say 120, actual 72” is not a mystery.
• How to check: the same content, the same peak cap, change only refresh, meter a white patch. Nits falling with hertz is the trade. If luminance holds, check whether duty cycle was raised — smear should have come back with it.

Related

• Same group: N1.12.1 Refresh rate jointly determines flicker visibility and motion continuity · N1.12.2 Low persistence trades shorter illumination for less smear · N1.12.3 When the eye pursues a moving object, smear is paint on the retina
• Nearby: N1.05 Motion-to-Photon Latency · N1.08 Headset Mass and Prolonged Wear
• Search terms: refresh rate · luminance · duty cycle