Blur Busters Forums

Forgive me if this's been asked before (first post here), but I've never really found a great answer to this besides something along the lines of "people don't like flicker." Is it something about LCDs and OLEDs that prevents them from being impulse-driven like a CRT? Of course, we can always wait until (hopefully) these non-CRTs finally achieve something like 960hz refresh rates, but won't we then also have to be able to run games at 960fps to see the benefits? I don't wanna have to have an RTX 3090 just to be able to run Half-Life 1 without motion blur. And if the problem really is that TV/monitor manufacturers think people will buy fewer monitors if they flicker, do they not know that flicker disappears to the average eye once you go above 60hz, especially when you hit 75hz? I'm typing this on a Dell M991 CRT right now running at 75hz and I don't perceive any flicker whatsoever. AT 60hz I do, though, but whatever. 120hz is becoming the norm for almost every display now, including TVs, and nobody's gonna see flicker at 120hz. So what's the big deal? Wouldn't everything be solved easily by just going back to impulse-driven? All the manufacturers seem to have just been putting a bunch of bandaids like ULMB and frame interpolation on this. I must be missing something here...

Excellent Educational Question

Be noted, scientifically:

1. ULMB is impulse driving.
2. Strobe backlights is impulse driving.
3. Strobe backlight is the only way to impulse-drive an LCD.
4. There are multiple strobe backlight brands (some better, some worse).
5. Remember I've seen thousands of displays. Strobe backlights can sometimes produce better motion quality than a CRT tube (i.e. Oculus Quest 2 LCD VR is among the best strobed LCDs I've ever seen).

Dead-Serious Question:
A. Have you ever seen a good strobe backlight before? As in REALLY compared 10 Or 100 models? I have. And some exceed CRT nowadays.

Strobe Backlights, in certain cases, can surpass CRT motion clarity
It is a false assumption that technologies similar to "ULMB" can't surpass CRT motion clarity. Sure, less than 1% of panels and strobe backlight settings can pull it off, but a well-tuned strobe backlight can produce a superior motion clarity experience. For example, a5hun's review of the Blur Busters Approved XG270 said that it produced better motion clarity than a Sony FQ900 at a very specific setting (119Hz-120Hz). Or the Oculus Quest 2 VR LCD or the Valve Index VR LCD. They are examples of strobe backlights that can (at times) surpass CRT motion clarity. Check out ApertureGrille XG270 Tests of 119 Hz Versus CRT and you will see how there are some fewer artifacts than CRT (strobe crosstalk much fainter than CRT phosphor ghosting), thanks to my Blur Busters Approved work on XG270 (and soon, a few models this year will have the Blur Busters Approved seal). Sure, it doesn't make the LCDs perfect nor surpass CRTs in all ways (imperfect blacks etc), but they reach nearly the pinnacle of strobe quality at some settings, as much as the LCD technology is able to permits.

CRTs impulse driving are also a humankind band-aid too
CRT impulse driving is also a humankind band-aid, because real world doesn't strobe. So I kind of resent the use of the word "Band-Aid" used out-of-context to ULMB but not to CRT flicker. Reading the Blur Busters Law: The Amazing Journey To Future 1000 Hz Displays, you will see how I say that Impulse driving of any kinds is a human-kind band-aid because the world has not developed analog motion displays to emulate real life. The artificial humankind invention of a "frame rate" in the world of real life being framerateless, meant that impulse driving is a good bandaid to eliminate motion blur of non-analog motion (aka series of stationary pictures to emulate real life, but real life isn't a series of stationarynesses).

Impulse Driving the LCD Panel via Software BFI
Now, if you wanted to impulse-drive the LCD panel instead of the backlight, it can be done via software black frame insertion, as seen at www.testufo.com/blackframes .... If you have a 240Hz display, view www.testufo.com/blackframes#count=4&bonusufo=1 ... That's software-based impulse-driving. But notice it never is as clear as a CRT. It's bottlenecked by LCD GtG and refresh cycle granularity time (e.g. "decay" can be never faster than a refresh cycle).

High Speed Videos: Understanding LCD Scanout in High Speed Video
This is how an LCD looks like in a high speed video: www.blurbusters.com/scanout .... Check these out and come back to this thread. In the past, LCD was very slow at refreshing. Here's a high speed video of an IPS LCD swapping 60 pictures per second:

See? You can't fit two GtGs per refresh cycle easily into that. Now, you've got faster/better LCDs that can do GtG almost as fast as an OLED, but still not quite fast enough to properly impulse-drive.

Theoretically Rolling Impulse Scan at the LCD Layer Level
Now, a custom LCD can theoretically multi-scan with an OFF scanout pass pass chasing the ON pass. I've got an OLED Custom Scan Patterns Thread that covers this. An example is this one. LCDs currently don't do this at the moment (Two GtG operations per refresh cycle is the only way to impulse drive at sub-refresh levels, if you don't impulse-drive via the backlight instead)


Impulse Driving Is Two GtGs Per Refresh Cycle
First GtG is black towards visible color
Second GtG is visible color back to black

Problem Trying To Impulse Drive Via LCD GtG: To Match CRT Clarity, Need 2 Super Fast GtG Per Hz
Pixel transitions (GtG) is a pixel fading from one color to the next. Even "1ms GtG" may be 10ms real world, see GtG vs MPRT: FAQ About Pixel Response. So to match CRT motion clarity with LCD-layer-based GtG impulse-driving you need GtG 100% to be 1ms for all possible color combinations, which is darn near nigh impossible. You need sub-millisecond GtG100% from black-to-color. Then another 2nd sub-millisecond GtG100% from color-back-to-black WITHIN the same refresh cycle to black out the pixel quickly.

The CRT Version of GtG
Even though we don't call it "GtG", CRTs have a metaphorical equivalent of GtG too:
The 1st GtG of a CRT for off-to-on is extremely fast (microseconds), the electron beam illuminating the phosphor dot.
The 2nd GtG for a CRT for on-to-off is the slower phosphor decay (still often decays 90% within a millisecond).

For LCDs, It Is Easier To Do Impulse Driving (Two GtGs Per Hz) Via Backlight
Backlights can turn on/off faster at much higher quality than trying to impulse-drive via LCD GtG. If you compare software BFI with hardware strobing, the quality of most hardware strobing can be superior on most panels.

More Reading -- References

1. Blur Busters Law: The Amazing Journey To Future 1000Hz Displays
2. Pixel Response FAQ: GtG Versus MPRT
3. High Speed Videos of LCD Refreshing
4. The Stroboscopic Effect Of Finite Frame Rates
5. Very Old 2014 Article: Electronics Hacking Creating A Strobe Backlight

For even more, study the great articles at Area51 -- www.blurbusters.com/area51 -- to gain a better understanding of displays.

Shopping Guides For Better LCD Impulse Driving (Via Strobe Backlight)
- CRT Nirvana Guide for DIsappointed CRT-to-LCD Upgraders
- Good News Everyone, 60 Hz Single Strobe Options

Chief Blur Buster wrote: ↑

22 Feb 2021, 01:31

Strobe Backlights, in certain cases, can surpass CRT motion clarity
It is a false assumption that technologies similar to "ULMB" can't surpass CRT motion clarity. Sure, less than 1% of panels and strobe backlight settings can pull it off, but a well-tuned strobe backlight can produce a superior motion clarity experience. For example, a5hun's review of the Blur Busters Approved XG270 said that it produced better motion clarity than a Sony FQ900 at a very specific setting (119Hz-120Hz). Or the Oculus Quest 2 VR LCD or the Valve Index VR LCD. They are examples of strobe backlights that can (at times) surpass CRT motion clarity. Check out ApertureGrille XG270 Tests of 119 Hz Versus CRT and you will see how there are some fewer artifacts than CRT (strobe crosstalk much fainter than CRT phosphor ghosting), thanks to my Blur Busters Approved work on XG270 (and soon, a few models this year will have the Blur Busters Approved seal). Sure, it doesn't make the LCDs perfect nor surpass CRTs in all ways (imperfect blacks etc), but they reach nearly the pinnacle of strobe quality at some settings, as much as the LCD technology is able to permits.

Just to point out some additional history. Callsignvega a member of the former HardOCP particularly popular member in hardforum. While OCP is gone, forums are up.

Back in 2015 using an eyefinity triple monitor in vertical mode. Apparently his monitors produced less motion blur than the mighty GDM-FW900 CRT. So even his CRTs probably a combination of either 2ms(500) or 3ms(333) phosphors and age of the CRT contributed to the LCDs producing less motion blur.

(Not all FW900s have the 2ms phosphors some had 3ms phosphors and to point additional history the SMPT-Phosphors were researched which reached 0.5ms(2000) clarity. I think if I recall these phosphors were left to professional equipment or large signage and never made it to the market since LCDs began dominating.)

While it's not probably 100% scientific to NOT question it. For a person using said screens to state a human response rather than a scientific inquiry is good.

For example you mentioned the HMD of Occulus Quest 2 or Valve index. Well funny enough back in 2015 with VERY specific monitor models the results were slightly below, at, or above CRT clarity. EVEN IF it's not 100% for some reason these monitors Callsignvega had were good ones.

It reminds me of certain things back in the first 120Hz LCDs like people stating the Viewsonic VX2268 was superior than the Samsung both in performance and in the fact the samsung was calibrated more towards the blue spectrum and had a colder color temperature while the VX2268 was warmer in color closer to 6500K. But non the less despite Samsung being good the Viewsonic panel for some reason performed better. I recall some people investigating this was back in what '09/'10 but the in depth review the Viewsonic outperformed the Samsung panel in some usage cases.

Whoa, wasn't expecting such a detailed response. I appreciate it.

First of all, I should've been more specific about impluse-driving a display panel vs. a backlight, but come to think of it, no I have not seen a backlight-strobed monitor, EXCEPT (now that I remember) for the PSVR, which IIRC is backlight-strobed in VR mode specifically, and I remember it looking as good as a CRT motion-wise, so you got me there.

Your section on impulse driving via LCD GtG - so is this the reason (along with flicker) why flat-panel displays went with sample-and-hold? Because the pixel-response time was just too long for impulse-driving the panel to have any sort of clarity during motion? If that's the case, then yeah now it makes sense to me.

Also, I get what you're saying about the "band-aid" term. I was just using that because, in my mind, the way CRTs work makes them (at least to our eyes) virtually inherently motion-blur free, whereas flat-panel displays have need all this extra tinkering in that regard since they came out. But I think I understand the situation a bit more now.

Anyway, thanks a lot for the response.

theplace1988 wrote: ↑

22 Feb 2021, 17:56

Your section on impulse driving via LCD GtG - so is this the reason (along with flicker) why flat-panel displays went with sample-and-hold? Because the pixel-response time was just too long for impulse-driving the panel to have any sort of clarity during motion? If that's the case, then yeah now it makes sense to me.

Correct. LCD GtG today (even the 0.5ms GtG ones) are still too slow for proper impulse-driving.

Also, another reason is that you need two GtG per refresh cycles. GtG behaviours are like kicking soccer balls (see GtG as Soccer Balls Metaphor as well as GtG is Physical Momentum in LCDs).

Metaphorically, faster GtG means you need a running start to kick the soccer ball harder, but if you have a higher refresh rate (even just to do two GtG's per refresh cycles), you have LESS time to give a running start to kick pixels harder.

So just even adding a second GtG pass per refresh cycle (to impulse drive ), slows down both GtGs because you have half the amount of time to give pixels a running kickstart (for faster GtG).

And since LCDs is Liquid Crystal Display, a GtG operation involves physical motion of molecules (which operate as defacto light valves/shutters to block/unblock light), which have momentum, which is hard to accelerate/decelerate.

Metaphorically, trying to design a car to go 500mph instead of 50mph, is tough -- getting things to move faster.