Does lightboost/ULMB make a big diff w/ fast response times

[ablaabla]

Suppose a monitor already has a very good response time (e.g. the Asus rog p278q with overdrive on has an average gtg time of < 3ms). So there is very little motion blur, but it is still there. Is this "good enough"? Is this at all perceptible? Certainly a photo will be able to show the difference, especially when there is high contrast between the updated image and the lingering afterimage. But would anyone actually notice the motion blur when using the monitor? Would someone at the 95% percentile of ability to perceive motion blur perceive such a minor amount of motion blur when using the monitor (especially at 144hz)?

Which brings me to my next question: does lightboost/ULMB make a noticeable difference on a monitor that already has an excellent response time (e.g. the p278q)?

EDIT: Also, I read somewhere that in the future there would be monitors that support simultaneous strobing + GSYNC. Does anyone have any details on release dates?

[sharknice]

Suppose a monitor already has a very good response time (e.g. the Asus rog p278q with overdrive on has an average gtg time of < 3ms). So there is very little motion blur, but it is still there. Is this "good enough"? Is this at all perceptible? Certainly a photo will be able to show the difference, especially when there is high contrast between the updated image and the lingering afterimage. But would anyone actually notice the motion blur when using the monitor? Would someone at the 95% percentile of ability to perceive motion blur perceive such a minor amount of motion blur when using the monitor (especially at 144hz)?

Which brings me to my next question: does lightboost/ULMB make a noticeable difference on a monitor that already has an excellent response time (e.g. the p278q)?

EDIT: Also, I read somewhere that in the future there would be monitors that support simultaneous strobing + GSYNC. Does anyone have any details on release dates?

It makes a huge difference and anyone can tell the difference.
Lower response times don't directly reduce motion blur, how long a pixel persists is the main factor. There is a pretty good explanation of why here: http://www.blurbusters.com/faq/oled-motion-blur/


So far there haven't been any announcements on Strobable GSYNC.

[RealNC]

Yep, correct. Low pixel response times do not reduce motion blur much. Even if you had a monitor with the impossible response time of 0 (infinitely fast), you'd still have motion blur at the refresh rates we have today. A sample-and-hold technology like LCD will always blur movement at low refresh rates.

If you had 1000Hz and also 1000FPS, which means 1ms frame persistence, then motion blur would be low. But as of right now, strobing is the only practical method of motion blur reduction, since even if we will have 1000Hz monitors, they're useless if you don't have 1000FPS content, which isn't gonna happen. Reaching 1ms frame persistence with, say 120Hz and a very good backlight is much more feasible then hoping for 1000Hz and 1000FPS, since the result would be the same (both would have 1ms frame persistence.)

Motion blur reduction will be mostly about advancing strobing technology (or other non sample-and-hold display methods). Chief mentioned somewhere that for really great results, we would need sub-1ms persistence. 0.5ms with 120Hz strobing for example *is* feasible. The equivalent without strobing would require 2000Hz/2000FPS, which as you can guess is extremely unlikely to happen.

Also, let's not forget that we actually *had* that in the past, even 20 years ago. CRT monitors did feature sub-1ms frame persistence, even at 60Hz. The industry just needs to bring that back into the LCD/OLED age by investing in backlight technology. I suspect the only reason they didn't, is that most people never complained about it. Not many consider motion blur to be a big issue.

[Black Octagon]

What I understand is that slow response times can cause motion blur. But if you have fast response times and want to reduce motion blur further, you need to get around the perceived motion blur that remains due to the way in which the human eye tracks motion in sample and hold displays. Hence ULMB, LightBoost and so on

[Chief Blur Buster]

Very massive difference.

For example, with 144Hz without strobing, it's impossible to successfully pass the TestUFO Panning Map Readability Test at any scrolling speeds 960 pixels/second or faster.

Click this link. The street labels are 100% readable on a CRT or strobed monitor (LightBoost, ULMB, Turbo240, BENQ Blur Reduction).

Milliseconds persistence is not the same thing as milliseconds response. Most 1ms 60Hz LCDs have 16.7ms of persistence. Persistence is pixel static/stationary time, and GtG response is pixel transition time. Both can muddy motion. Even when you have 0ms instant response, there still motion blur caused by the sample-and-hold effect. A refresh is continuously shining. As you track your eyes, your eyes are in a different position at the ending of a refresh cycle, than at the beginning of a refresh cycle. Even if refreshes instantly changed (0ms) to the next frame, the act of eye motion across static frames of a non-strobed LCD refresh, causes the static frame to be blurred across your retinas.

Here is an excellent animation of motion blur that is NOT CAUSED by LCD response. The motion you see in this motion test, is not caused by LCD response limitations. It is caused by the sample-and-hold effect (aka "persistence").

1. Look at the stationary UFO
2. Look at the moving UFO

The optical illusion you see, is an optical illusion that I invented, to demonstrate motion blur caused by persistence (eye-tracking). This type of motion blur is NOT caused by LCD response. This motion blur optial illusion becomes sharper (checkerboard becomes clearer as response time goes to 0ms), since the motion blur optical illusion is mathematically optimized to become a more perfect checkerboard, the closer LCD response time approaches 0.

This also manifests itself in the TestUFO Panning Map Readability Test as well.

[flood]

What I understand is that slow response times can cause motion blur. But if you have fast response times and want to reduce motion blur further, you need to get around the perceived motion blur that remains due to the way in which the human eye tracks motion in sample and hold displays. Hence ULMB, LightBoost and so on

one approximate way to think about it is

motion blur = persistence + response time

ofc you can't characterize response time by a single number, but it helps put into perspective, for instance, how little 1ms vs 2ms gtg matters compared to, for instance, unstrobed 60hz (16.7ms persistence) vs 144hz (6.94ms persistence)

because of this, a strobed 85hz crt would exhibit less motion blur than a 240hz lcd with instantaneous response time

[Edmond]

I would take that 240hz gsynced TN LCD over anything strobed right now.

I love the result of strobing, but you also get flicker, and the same tearing vs stutter and input lag crap as always.
So i firmly stand by gsync > all.

The next step would be to combine gsync with OLED to make response times irrelevant entirely. I hear motion is quite more clear on a 60hz OLED than on 60hz LCD even, just because of the sub millisecond response times. Also, thx to OLED, gsyncs minimum framerate could be 0 instead of 30, that would remove that flicker on loading screens and when framerate drops very low... a true fps=hz, with a max cap.

And the next step is to raise the max refresh rate as high as possible to achieve lower and lower persistence. Because, just tasting low persistence at the cost of other severe quality aspects is not worth it, IMO. If playing current games is a goal that is... if you play 10 yearold games only, fuck it, go full strobage.

[Falkentyne]

I Think you're sort of wrong here.

Instant response time would NOT give you no motion blur.
You would have the SAME motion blur you have now.
What you would have is--no overdrive artifacts. Because you wouldn't need overdrive.

The only to have no motion blur on a 0 response time (0ms instantaneous) panel, is to be running at 250 or higher hz (250hz= 0.4ms) and have the framerate at that refresh rate with vsync (or close to it if you're using gsync). 500hz would be 0.2ms and 1000hz would be 0.1ms.

And to get that silky smooth 0.1ms, you would need 1000 fps.
No game in recent memory can run at such a FPS anyway.

So right now we can get that 0.1ms through storbing and black frame insertion: 100hz + strobe duty 010 on the Benq Z series blur reduction=0.1ms. But you have overdrive ghosting since the panel doesn't have a 0ms response time.

Now, if it did.......guess what?
You got yourselves a strobed CRT

Strobing has been in CRT's since the dawn of television. But LCD panels were never designed with strobing as an inherent technology like CRT's were. A completely new panel, designed purely to be run strobed, would have to be designed for this, but you'd need a 0 ms response across all pixel transitions, so we can throw overdrive away forever.

(Am I right, Chief?)

[Chief Blur Buster]

The math is incorrect.
On a scientifically "perfect" sample-and-hold display:

Persistence = refresh visibility time

120Hz sample-and-hold = 1/120 = 8.33ms persistence
240Hz sample-and-hold = 1/240 = 4.16ms persistence
1000Hz sample-and-hold = 1/1000 = 1ms persistence
10,000Hz sample-and-hold = 1/10,000 = 0.1ms persistence

Typical strobe backlights are:
LightBoost 100% = 2.4ms strobe flash = 2.4ms persistence = ~1/400sec = requires 400fps@400Hz flickerfree to match
LightBoost 10% = 1.4ms strobe flash = 1.4ms persistence = ~1/700sec = requires 700fps@700Hz flickerfree to match
ULMB = 2ms strobe flash = 2ms persistence = 1/500sec = requires 500fps@500Hz flickerfree to match

Now when you go with the calibrateable strobe backlights with adjustable strobe length (persistence), such as BENQ Blur Reduction and some newer ULMB monitors, you can go to very low values such as 0.5ms (and less)
0.5ms = 1/2000 = requires 2000Hz to match motion blur
0.2ms = 1/5000 = requires 5000Hz to match motion blur

[Chief Blur Buster]

The only to have no motion blur on a 0 response time (0ms instantaneous) panel, is to be running at 250 or higher hz

Not quite that simple.

Although vastly less motion blur than 144Hz, and many people may prefer 250Hz without a strobe backlight, there will still be visible motion blur. 250fps@250Hz will be the same amount of motion blur as a strobe backlight using strobe flash length of 1/250sec = 4ms. That's worse than LightBoost 100%. It would begin to almost fail the TestUFO Panning Map Test, as an example.

Persistence is refresh visibility time.

Therefore,
Persistence on strobe backlights is the strobe flash time.
Persistence on flickerfree display is the length of the refresh cycle.
Persistence on a CRT is the persistence of the phosphor (amount of time it glows)

You can't simply choose specific numbers like "250", as it depends on the use case. An everyday user running slow games won't see the benefits of a 120Hz unstrobed monitor (8.3ms persistence) and see no motion blur. While a hard-core gamer who loves CRT zero motion blur, will still be disgusted by the motion blur of a 250Hz unstrobed monitor.

In the real world, you have the fade-to-black (e.g. CRT) or blend-to-next-refresh effect (e.g. GtG transition), so slower response will "add" to the persistence, in a kind of "motion blur = persistence + GtG response" .... Of course, not quite exactly mathematically that simple, due to curves, but you get the idea that slower GtG response will add to persistence, while instant zero GtG response won't eliminate motion blur because of persistence.

So once you got zero GtG, the only way to halve motion blur is to halve the frame visibility time.
Halving motion blur is only possible via:
1. Flickerfree method of halving motion blur: Double the refreshrate(and framerate, either real or interpolated)
2. Flicker method of halving motion blur: Have each pixel spend spend half of the refresh cycles displaying nothing between refreshes (e.g. black frame, CRT scan, OLED rolling scan, LCD backlight strobe flash length).

Either way, that halves refresh visibility time, so that's half persistence as before, and thus halves motion blur (assuming GtG is no longer the limiting factor). In modern LCDs, GtG is no longer the limiting factor of motion blur, since GtG is less than a tenth of the cause of the motion blur on a modern 1ms 60Hz LCD (1ms GtG versus 1/60sec persistence = 16.7ms persistence)