Monitor Backlight Strobing (ULMB / DyAc) Explained: Why Fast Motion Gets Sharper but Dimmer — MPRT, Strobe Crosstalk & the VRR Conflict

｜ HK GPU Fix

重點摘要

Backlight strobing shortens the time each frame is visible, so fast-moving objects look sharper even though overall brightness falls and flicker can appear. This guide explains MPRT, strobe crosstalk, and why strobing usually cannot run together with variable refresh rate before you switch it on.

Two Different Kinds of Motion Blur

Motion blur on a display is not one phenomenon, and the confusion between its kinds is responsible for most of the disappointment people feel after changing a display setting and seeing no improvement. There are two main layers, and they have different causes, different cures, and different names.

The first layer is pixel response blur. A liquid crystal pixel takes time to change from one colour to another. If the pixel has not finished changing when the next frame arrives, the eye sees a smear trailing behind moving edges. This is the layer governed by grey-to-grey response time, and it is the layer addressed by overdrive, which pushes pixels harder to make them change faster. Overdrive is covered in depth in article #214, and this article does not repeat it, because the two mechanisms must not be confused.

The second layer is eye-tracking blur, and it is the subtler one. Most modern flat displays are sample-and-hold displays: each frame is shown continuously until the next frame replaces it. The image does not flash and vanish the way an old cathode-ray tube image did; it persists. When the eye follows a moving object across such a display, the eye moves smoothly while the object jumps from one fixed position to the next at each frame. During each frame's hold time, the eye has already moved on, so the object's image is smeared across the retina. The result is blur that persists even if the pixels themselves switch instantly. Faster pixels cannot fix this layer, because the cause is not slow pixels at all: it is the combination of a held image and a moving eye.

Raising the refresh rate shortens each frame's hold time, which reduces eye-tracking blur, but it does not remove it. At any finite refresh rate, each frame is still held for the full frame interval, and the eye still smears it. This is why a fast display can still look soft in fast motion, and why a further technique is needed for players who want the sharpest possible moving image.

How Backlight Strobing Works

Backlight strobing attacks the eye-tracking layer directly, by shortening the time each frame is actually visible. Instead of leaving the backlight on for the whole frame interval, the display flashes the backlight on for a brief slice of each frame and leaves it off for the rest. The pixels still change behind the scenes, but the eye only sees the frame during the brief lit window, after the pixels have settled. The smeared part of the story, the pixel transition itself, happens in darkness and is never seen.

The effect is conceptually close to how an old cathode-ray tube behaved. A tube's phosphor flashed briefly and then faded, so each frame was visible only for a short moment, and the eye never had a held image to smear. Strobing recreates that short-visibility behaviour on a modern panel. The moving object is seen as a series of brief, sharp snapshots rather than a continuously held image, and the eye-tracking smear collapses accordingly.

This is also where the idea of MPRT, moving picture response time, comes in. MPRT describes how long a moving image is effectively visible, which in a strobed display is close to the backlight's lit time. In a non-strobed hold-type display, the effective visible time cannot be shorter than one full frame interval, because the frame is shown for the whole interval. With strobing, the effective visible time can be shorter than the pixel transition time itself, because the transition is hidden in the dark period. That distinction is the whole point of the technique, and it is also the source of the most common specification confusion, addressed below.

Strobing goes by many brand names, and recognising them as one family prevents a great deal of marketing confusion. NVIDIA's ULMB and its later ULMB 2, BenQ's DyAc line, ASUS's ELMB, ViewSonic's PureXP, and Gigabyte's Aim Stabilizer are all implementations of the same underlying idea: brief backlight illumination per frame, or black frame insertion, to shorten visible time. The names differ, the menus differ, and the tuning differs, but the mechanism is shared. No single brand name owns the technique, and no brand name by itself guarantees a better result, because the outcome depends on the panel, the refresh rate, the strobe timing, and the brightness trade discussed below.

A small number of newer designs attempt to combine strobing with variable refresh rate, a combination that traditional strobing cannot offer. These remain the exception rather than the rule, and they apply to specific models. Players should judge by what their own display's menu currently offers, not by the existence of a newer design elsewhere. The general rule for the overwhelming majority of displays is the one this guide assumes: strobing and variable refresh are alternatives, not companions.

MPRT Versus GtG: Reading Specifications Honestly

Display specifications invite a particular misunderstanding, and it is worth dismantling carefully, because it drives poor buying decisions. Two response figures appear on boxes and product pages, and they measure different things.

Grey-to-grey time, GtG, measures how long a pixel takes to change from one shade to another. It describes the pixel layer of blur. Overdrive exists to improve this figure, and a fast GtG figure is genuinely desirable, but it says nothing about the eye-tracking layer. A panel with fast pixels on a hold-type display still holds each frame for the full frame interval, and the eye still smears it.

MPRT measures how long a moving image is effectively visible. On a non-strobed hold-type display, MPRT is at least one frame interval, whatever the GtG figure says. On a strobed display, MPRT approaches the backlight lit time, which can be shorter than GtG, because the pixel transition is hidden in darkness. A strobed display can therefore show a very short MPRT figure alongside a modest GtG figure, and both figures can be true at once, describing different layers.

The trap is comparing a single millisecond figure across displays without knowing which layer it describes or how it was measured. A headline figure labelled as MPRT is not comparable to a headline figure labelled as GtG, and neither figure, taken alone, predicts how a display will look in motion. Measurement methods differ between makers and reviewers, test patterns differ, and the strobe settings active during measurement differ. The honest way to read specifications is as a starting point for shortlisting, not as a verdict. The verdict comes from looking at the display in motion, with strobing on and off, under the conditions in which it will actually be used.

It follows that a shorter MPRT figure is not automatically better in every respect. The figure says nothing about how much brightness was sacrificed to achieve it, nothing about strobe crosstalk, nothing about the range of refresh rates over which strobing works well, and nothing about how the display feels to a particular pair of eyes over a long session. All of those factors belong in the judgement, and the sections below take them in turn.

Diagnosis: Which Blur Are You Actually Seeing?

Before switching strobing on, it is worth establishing which layer of blur is actually bothering you, because strobing only helps one of them.

Start with a still image. If a static scene already shows trailing edges, ghosting, or smearing around objects that are not moving, the problem is unlikely to be eye-tracking blur. Still-image artefacts point first to overdrive behaviour or to rendering-layer artefacts, and those should be investigated through article #214 before strobing is considered. Strobing a still image achieves nothing, because there is no eye movement to smear anything.

Next, watch a moving object against a detailed background. If the scene is sharp when still and softens only once objects start moving, and especially if the softening follows the eye as it tracks a moving target, the eye-tracking layer is a plausible contributor, and strobing is a reasonable experiment.

Then consider how you actually play. Strobing improves the clarity of objects the eye tracks while they move. If your play style involves long periods of staring at a fixed central point with little eye movement, the benefit will be much smaller than for a player whose eyes constantly sweep across fast-moving action. The technique's value depends on viewing behaviour, not only on the display, and a player whose style does not involve much tracking should not be surprised by a modest result.

Finally, turn off in-game motion blur before testing. Many games add their own blur effect to moving images, and that effect will mix with the display's behaviour in a way that makes attribution impossible. A test conducted with the game's blur still on cannot tell you whether the display's strobing helped, because two blur mechanisms are active at once. Switch the game's effect off, test the display alone, and only then decide whether either is wanted.

The On/Off Comparison: Three Things to Judge Together

A fair strobing comparison judges three things at once, not one. Players who judge only sharpness will conclude that strobing is a pure win, and players who judge only brightness will conclude it is a pure loss. Both judgements are incomplete.

Use a fixed refresh rate and a frame rate that stays close to it, in the same game and the same scene, with the game's own motion blur off. Switch strobing on and off without changing anything else. Then judge sharpness first: do fast-moving objects, such as a strafing opponent or a scrolling landscape, resolve more clearly, with less smear behind their edges? Judge brightness second: how much dimmer is the whole image, and is the dimmer image still comfortable in the room where you play? Judge comfort third: does the image flicker noticeably, and do your eyes tire faster than with strobing off?

Brightness deserves emphasis, because it is not a minor side effect. Since the backlight spends most of each frame switched off, total light output falls substantially. In a bright room, the strobed image may simply be too dim to be usable. In a dark room, the same image may be perfectly comfortable. The judgement is therefore inseparable from the room, and a review conducted in a dark studio does not transfer to a sunlit desk.

Flicker deserves equal emphasis. Strobing is, by design, a flashing light. At higher strobe frequencies the flicker is less noticeable to most people, but at lower frequencies it becomes perceptible, and some people are markedly more sensitive to it than others. Anyone who notices flicker, eye strain, or discomfort should switch strobing off without hesitation. No sharpness gain is worth persisting with a display that causes discomfort, and the correct response to discomfort is not to endure it in the hope of adapting.

Strobe Crosstalk: The Ghost That Strobing Itself Creates

Strobing removes one artefact and can introduce another, and recognising the new artefact is essential to using the technique well. Strobe crosstalk appears as a faint double image or ghost trailing a moving object, most often near the top and bottom edges of the screen, with the centre usually cleaner. It is not the same as overdrive ghosting, and confusing the two leads to the wrong fix.

The cause is timing. Strobing assumes that all pixels have finished changing before the backlight flashes. In practice, pixels do not all change at the same instant: the panel updates from one edge to the other, so at the moment of the flash, some rows have settled while others are still mid-transition. Rows caught mid-transition are illuminated while showing a half-changed colour, and the eye sees a faint second image offset from the main one. That is crosstalk: a ghost created by illuminating pixels before they finished changing.

Crosstalk is typically worst at the screen edges and mildest at the centre, because the centre rows are closest to the ideal timing point. It also tends to worsen at the highest refresh rates, because the dark interval available for pixel transitions shrinks as the frame interval shrinks, leaving less margin for slow rows. This produces the counter-intuitive result that lowering the refresh rate can sometimes produce a cleaner strobed image than the maximum refresh rate: the longer frame interval gives pixels more time to settle before the flash, and the centre-clean region widens.

If crosstalk is severe enough to interfere with judgement, the remedies are to try a different strobe level where the display offers one, to lower the refresh rate to a value where the panel has more settling time, or to switch strobing off. Raising the strobe level or forcing the maximum refresh rate in the hope that more will fix it is the wrong direction. Crosstalk is a timing problem, and timing problems are solved by giving the panel more time, not less.

A table helps keep the four look-alike artefacts apart, because each one calls for a different remedy and each one is regularly mistaken for another.

A second table summarises the trade that strobing makes, so that the decision can be taken with eyes open rather than on the strength of a single benefit.

The Variable Refresh Conflict

The most important practical limitation of traditional backlight strobing is that it generally cannot run at the same time as variable refresh rate. This is not an arbitrary product restriction; it follows from how the two techniques handle timing.

Variable refresh works by varying the interval between frames: the display waits for each finished frame and refreshes when it arrives, so the frame interval changes with the game's frame rate. Strobing works by flashing the backlight at a fixed rhythm tied to a fixed frame interval: the flash must land at a precise point in each frame's cycle, after pixels have settled. If the frame interval keeps changing, there is no fixed point at which to schedule the flash, and the timing that hides pixel transitions breaks down. The two timing models are therefore in tension.

In practice, enabling variable refresh on most displays greys out the strobing option or switches it off automatically, and enabling strobing disables variable refresh. The player must choose: tear-free, variable timing with softer motion, or fixed timing with sharper motion. Neither choice is wrong in itself; the mistake is expecting both at once and then concluding that one of the features is broken when the menu refuses to combine them.

A small number of newer designs attempt to reconcile the two, by adapting the strobe timing to the variable interval. These are specific to particular models, and their availability should be judged by the display's own menu, not assumed. For the great majority of displays, the honest statement is the simple one: strobing and variable refresh are alternatives, and the choice between them is a real trade-off to be made per game and per preference, not a setting to be switched on and forgotten.

This is also the point at which the two kinds of flicker must be kept apart. Variable refresh can produce brightness flicker in dark scenes, because the display's brightness behaviour shifts as the refresh rate varies. That flicker has a different cause from strobing flicker, which is the deliberate flashing of the backlight, and the two are addressed differently. Readers who want the variable-refresh flicker story in full should read article #205, which explains its gamma-related cause without mixing it into the strobing account.

Frame Rate Must Match the Refresh Rate

Strobing imposes a discipline on frame rate that non-strobed play does not. The technique works best when the frame rate stays close to the refresh rate, so that each strobe flash corresponds to one fresh frame. When the frame rate falls well below the strobe frequency, the display repeats frames, and the eye sees the same object position flashed more than once per movement step. The result is multiple images of the same object, a juddering, multi-ghost appearance that is worse than the blur strobing was meant to remove.

Unstable frame rates cause a related problem. If the frame rate wanders, the relationship between strobe flashes and fresh frames wanders with it, and the sharpness benefit varies from moment to moment. A game that cannot sustain a frame rate near the chosen refresh rate is a poor candidate for strobing at that refresh rate, and the practical remedy is sometimes to lower the refresh rate to a value the game can sustain, rather than to insist on the maximum. A lower refresh rate with a matching frame rate and clean strobe timing will look sharper than a maximum refresh rate with mismatched frames and heavy crosstalk.

This requirement also interacts with the choice of game. A lightweight competitive title that holds a steady high frame rate is a natural strobing candidate. A demanding single-player title whose frame rate swings widely is not, even if its motion would benefit in principle. The technique rewards stable performance, and judging it in an unstable game will understate what it can do in a stable one.

Boundary Cases: Brightness, Eyes, OLED, and Other Limits

Several boundary cases deserve explicit treatment, because each one turns a reasonable experiment into a poor experience if it is ignored.

Brightness is the first. The dimming caused by strobing is inherent, not a defect, and it cannot be fully compensated by raising the brightness control, because the backlight is off for much of the frame regardless of how bright it is during its lit slice. A display used in a bright room, or by a player who prefers a bright image, may simply be unsuitable for strobing, and that is a legitimate conclusion rather than a failure of technique.

Flicker sensitivity is the second. At lower strobe frequencies, flicker becomes noticeable to many people, and a minority are strongly sensitive even at higher frequencies. Sensitivity is a property of the person, not a flaw in their technique, and the advice is unconditional: if strobing causes discomfort, eye strain, or headache, switch it off. The guide makes no medical claim about strobing and offers no assurance about safety; it reports only that some people notice flicker and discomfort, and that discomfort is a sufficient reason to stop.

OLED displays are the third boundary case, and they are often misunderstood. OLED pixels switch extremely fast, so the pixel-response layer of blur is minimal on OLED. But OLED is still a hold-type display: each frame is shown continuously until replaced, so eye-tracking blur still applies in principle. OLED black-frame insertion and strobing-like techniques exist, but they are not the same mechanism as LCD backlight strobing, because there is no backlight to flash. LCD strobing explanations should not be transferred wholesale to OLED, and OLED buyers should judge OLED motion features on their own terms.

In-game motion blur is the fourth. As noted in the diagnosis section, a game's own blur effect will mix with display strobing and confuse any comparison. Switch the game's effect off before testing the display, and decide about each effect separately.

Finally, panel uniformity matters. Crosstalk varies across the screen, and some panels show a much cleaner centre region than others. Two displays with the same headline specification can therefore look quite different with strobing on, and a specification sheet cannot capture that difference. Where possible, the judgement should be made by looking at the actual display, in the actual room, in the actual game.

Common Myths, Corrected

Several myths about strobing recur, and each one leads to a specific mistake.

"A smaller GtG figure always means sharper motion." It does not. GtG covers only the pixel layer. The eye-tracking layer depends on how long each frame is visible, and a fast-pixel hold-type display still smears in the eye. Both layers must be considered.

"Strobing is the same as variable refresh." It is not. Strobing shortens visible time at a fixed rhythm; variable refresh varies the refresh moment to match frame delivery. The two usually cannot run together, and enabling one is not a substitute for the other.

"Strobe crosstalk is the same as overdrive ghosting." It is not. Crosstalk comes from illuminating pixels before they finish changing; overdrive ghosting comes from pushing pixels past their target. The fixes differ: crosstalk is eased by more settling time, overdrive ghosting by a lower overdrive level.

"A smaller MPRT figure is always better, regardless of anything else." It is not. MPRT figures depend on measurement method and strobe settings, and they say nothing about brightness, crosstalk, refresh-rate range, or eye comfort. A figure without those companions is not a verdict.

"The highest refresh rate always gives the cleanest strobed image." It does not. At the highest refresh rate, the dark interval for pixel settling is shortest, and crosstalk can worsen. Sometimes a lower refresh rate, matched to a sustainable frame rate, produces the cleaner result.

When Strobing Is Worth It, and When It Is Not

Strobing earns its place for a fairly specific player: someone who plays fast-moving games at a stable frame rate close to the refresh rate, in a room where a dimmer image is comfortable, on a display whose strobe timing is clean across most of the screen, and who does not notice flicker. For that player, the sharpness gain is real and immediately visible, and the trade-offs are acceptable.

It is a poor choice for a player in a bright room, for a game with a wandering frame rate, for a player who wants variable refresh at the same time, or for anyone who notices flicker or eye strain. In those situations the honest answer is to leave strobing off, and no spec-sheet figure should override what the eyes report.

The broader lesson mirrors the one that applies to display settings generally: every motion-clarity technique addresses one layer of a multi-layer problem, and no single setting addresses them all. Pixel response, eye-tracking hold time, strobe timing, and variable refresh are separate mechanisms with separate trade-offs. The player who can name which layer is bothering them, and test one change at a time, will arrive at a better setup than the player who enables every clarity feature at once and hopes the combination sorts itself out.

Frequently Asked Questions

Why does my display look dimmer with ULMB or DyAc switched on?

Because the backlight is switched off for much of each frame. Only a brief lit slice per frame reaches your eyes, so total light output falls. The dimming is inherent to the technique, not a fault, and in a bright room it may make strobing impractical.

Can I use backlight strobing together with G-Sync or FreeSync?

On most displays, no. Traditional strobing needs a fixed frame rhythm so the flash can be timed after pixels settle, while variable refresh varies the frame interval. Enabling one usually disables or greys out the other. A small number of newer designs attempt to combine them, but availability depends on the specific model, so check your own display's menu.

What is strobe crosstalk, and how is it different from overdrive ghosting?

Strobe crosstalk is a faint double image caused by the backlight flashing before some pixel rows finish changing, usually stronger near the screen edges. Overdrive ghosting is a halo caused by pushing pixels past their target colour. Crosstalk is eased by giving pixels more settling time, such as a lower refresh rate, while overdrive ghosting is eased by lowering the overdrive level.

Is a smaller MPRT figure always better?

No. MPRT and GtG measure different layers, and MPRT figures depend on measurement method and strobe settings. A headline MPRT figure says nothing about brightness loss, crosstalk, refresh-rate range, or eye comfort, so it should not be used alone to choose a display.

Why does strobing look worse at the highest refresh rate?

Because the frame interval is shortest at the highest refresh rate, leaving the least dark time for pixels to settle before the flash. Crosstalk can therefore worsen. Lowering the refresh rate to a value the game can sustain sometimes gives a cleaner strobed image.

Does strobing help if my game already has motion blur switched on?

The two effects will mix and confuse any comparison, so switch the game's motion blur off before testing strobing. Judge the display technique and the game effect separately, and only then decide whether either is wanted.

Should I keep strobing on if I notice flicker or eye strain?

No. Switch it off. Flicker sensitivity varies by person, and discomfort is a sufficient reason to stop. No sharpness gain justifies persisting with a display setting that causes eye strain.

A Final Note

Backlight strobing is one of the few display techniques whose benefit is instantly visible and whose cost is equally visible. Sharper motion, dimmer image, possible flicker, no variable refresh: the trade is stated plainly, and the choice is personal. Test it in your own game, in your own room, judging sharpness, brightness, and comfort together, and let the result, rather than a specification figure, decide.

For help choosing display settings for a particular game, or for a check of whether motion softness is coming from pixel response, eye-tracking hold time, or strobe timing, readers are welcome to contact the shop at Room 1209, Dali Plaza, Shek Mun, by phone on 2186 8010, or by WhatsApp on 6558 6806.