upcoming display technologies that aren't sample-and-hold?

Everything about displays and monitors. 120Hz, 144Hz, 240Hz, 4K, 1440p, input lag, display shopping, monitor purchase decisions, compare, versus, debate, and more. Questions? Just ask!
willrs
Posts: 19
Joined: 04 Nov 2021, 18:12

upcoming display technologies that aren't sample-and-hold?

Post by willrs » 19 Jun 2022, 12:42

Like Mark has said, BFI is a band aid. Yes it helps but I don't know if it will ever fully eliminate motion blur and I agree it's not ergonomic (especially on OLED, I've seen BFI been a bit too flickery at times. BFI seems more ergonomic on LCD's). I think it's fair to say that if you care about motion blur, sample-and-hold displays are just inherently flawed, it's completely unrealistic to expect devices to record video at 480 or 960 frames per second, or video game consoles to render 480 or 960 frames per second. (And what about existing content?) So, my question is, is there any non-sample and hold displays being developed right now? And also, surely people involved in these decisions knew that sample-and-hold has an inherent motion blur flaw, why invest in them so much?

valeriy l14
Posts: 30
Joined: 21 Mar 2021, 11:36

Re: upcoming display technologies that aren't sample-and-hold?

Post by valeriy l14 » 22 Jun 2022, 15:51

willrs wrote:
19 Jun 2022, 12:42
Like Mark has said, BFI is a band aid. Yes it helps but I don't know if it will ever fully eliminate motion blur and I agree it's not ergonomic (especially on OLED, I've seen BFI been a bit too flickery at times. BFI seems more ergonomic on LCD's). I think it's fair to say that if you care about motion blur, sample-and-hold displays are just inherently flawed, it's completely unrealistic to expect devices to record video at 480 or 960 frames per second, or video game consoles to render 480 or 960 frames per second. (And what about existing content?) So, my question is, is there any non-sample and hold displays being developed right now? And also, surely people involved in these decisions knew that sample-and-hold has an inherent motion blur flaw, why invest in them so much?
Hell, you just need the display to be updated at a frequency 2 times higher than its horizontal resolution (because it is unlikely that any gamer will be able to pan the entire screen in 250ms) + the computer must give the same fps (those for 1920x1080 you need the refresh rate of the display is 3840 Hz, for 4k it is 7680, for 8k 15360 (and so on until you can no longer distinguish between a single pixel, or ego offset)) if you want to get a crystal clear image in dynamics in 99.99% of cases.

Think of it this way. You move the mouse cursor and it crosses the whole screen in 1 second (let's say we have a 2560x1440 display). In theory, a fat trail should follow the cursor like in real life when we quickly wave the lights (especially at night) + the cursor itself should remain statistically clean, but alas, since our display has a refresh rate of only 120 Hz, it tries to plug the gap of 2560 pixels to 120 pixels . And what's going on? Correctly ! Strobe + sample and hold effect. Now you can see that every object in motion has a 21.3 pixel blurry trail behind it! Now imagine when you have the whole scene moving at such a speed .... horror.

If we are talking about bfi (for convenience, ulmb ), then there the blur level is determined by the formula frame time divided by the time the backlight is on. Example we have a 120 hz display , it has a frame time of 8.3 ms but if we turn the ulmb on so that the backlight is lit only 12.5% ​​of the time then we get 8.3 : 8 = ~ 1 ms of pixel visibility (consider this as analog
1000 Hz) . But alas, the lower the backlight visibility time, the dimmer the display.


And the fact that Mark has been promoting 1000 Hz for more than a year is (in my opinion) due to the fact that at this frequency you can turn on the Ulmb and not worry that the eyes will not pop out of their sockets from pain after 10 minutes. Well, there are 2 more options: he believes that this frequency is enough for a comfortable + - feeling of a clear picture in motion, since at 500 ms we will have a 1.06 mm loop from the object (regardless of what screen resolution you have, be it 1080p or 64k or even more) and not 17.7 mm at 60 Hz

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: upcoming display technologies that aren't sample-and-hold?

Post by Chief Blur Buster » 22 Jun 2022, 21:26

valeriy l14 wrote:
22 Jun 2022, 15:51
And the fact that Mark has been promoting 1000 Hz for more than a year is (in my opinion) due to the fact that at this frequency you can turn on the Ulmb and not worry that the eyes will not pop out of their sockets from pain after 10 minutes. Well, there are 2 more options: he believes that this frequency is enough for a comfortable + - feeling of a clear picture in motion, since at 500 ms we will have a 1.06 mm loop from the object (regardless of what screen resolution you have, be it 1080p or 64k or even more) and not 17.7 mm at 60 Hz
willrs wrote:
19 Jun 2022, 12:42
Like Mark has said, BFI is a band aid. Yes it helps but I don't know if it will ever fully eliminate motion blur and I agree it's not ergonomic (especially on OLED, I've seen BFI been a bit too flickery at times. BFI seems more ergonomic on LCD's). I think it's fair to say that if you care about motion blur, sample-and-hold displays are just inherently flawed, it's completely unrealistic to expect devices to record video at 480 or 960 frames per second, or video game consoles to render 480 or 960 frames per second. (And what about existing content?) So, my question is, is there any non-sample and hold displays being developed right now? And also, surely people involved in these decisions knew that sample-and-hold has an inherent motion blur flaw, why invest in them so much?
BFI will still be needed for decades, indeed.

Good BFI can still benefit from 1000Hz+ displays of the future, too. The holy grail would be display emulators (CRT electron beam emulator, plasma display emulator, etc) that piggyback on the sheer Hz for accurate temporal display emulation (including low blur, softer rolling flicker, phosphor decay, etc) and does whatever temporal technology is needed to comfortably reduce display motion blur.

Emulating a CRT electron beam in software is one of my long-term aspirations too as well -- but it requires sheer Hz. e.g. using 16 digital refresh cycle to emulate 1 CRT refresh cycle, etc. (More Hz is better -- e.g. 4000Hz+ -- but a hell lot of display emulators can be begin to be written temporally accurately, as refresh rates ever get higher). The more Hz, the more accurately a retro display can be temporally simulated. I've written about this many times, especially in Area51 Forum and elsewhere.

Some of the efforts I do is to try and speed up the refresh rate race a bit, since geometric upgrades are needed long-term human visible benefit. The diminishing curve of returns is a tough wall, e.g. 60 -> 120 -> 240 -> 480 -> 960 ... As most users need to upgrade by 2x-4x to see human visible differences.

Things like 240Hz-vs-360Hz, normally a 1.5x differential, is only a 1.1x differential due to extra blur from non-zero LCD GtG and blur from high-frequency jitter (From all sources -- mouse / game / microstutter). So yes, we are hitting limitations caused by things like mouse poll rate and the game engine.

Now that being said, more Hz in a display can benefit lower Hz for many reasons, especially if it's from a near-zero-GtG display tech such as a future 240 Hz or 360 Hz OLED. By the end of the decade, most high Hz displays will likely become OLED instead of LCD, and paves a clearer path (pun intended) to 1000Hz.

There is the frame rate amplification technology (e.g. future variants of DLSS that can do 4x-10x frame rate increases) which will be something achievable within the next decade.

But we don't need that necessarily for low frame rate content -- however ultra high Hz can help lower Hz massively in many ways -- even consoles. As you can have temporal BFI filters that accurately simulates a retro display of your choice in motion clarity / phosphor decay / plasma noise / DLP color wheel / whatever artifacts you want or don't want. The more Hz, the more temporally accurate a retro display can be emulated. Multi-KHz displays can pretty much (to required retina leagues for most humans) emulate the temporal look and feel of a CRT of your choice, e.g. using 10, 20, 50 or even 100 refresh cycles to emulate 1 legacy refresh cycle temporally accurately.

Yes, extra brightness is needed for that (but that's where HDR comes useful; to produce the necessary windowed brightness surge headroom for things like CRT electron beam simulators). The spot is super bright in a high speed video. But did you know playing back a 1000fps high speed video of a CRT tube, back onto a 1000Hz display in realtime (not slomo), fairly accurately temporally 'simulates' the original CRT itself -- phosphor fade, rolling scan, zero blur included? Now, we just have to do it in software.

Based on what I've visually seen, this is essentially a BFI Holy Grail. Whether done as a shader app in a Windows indirect display driver, or as a video processor box (like HD Fury with the necessary FPGA logic for a CRT electron beam simulator). This will be something that will incubate over the next many years, but as a major Blur Busters mission, we are pushing the refresh rate race.

Also, 1000fps+ 1000Hz+ also makes VRR kinda obsolete because the ultrafine pixel refresh behavior behaves as defacto per-pixel VRR. You can have 24fps, 25fps, 48fps, 50fps, 59.94fps and 60fps video playing without visible pulldown judder, because of the ultrafine 1ms refresh cycle opportunities -- even side by side. So 1000Hz displays also benefits low frame rates without BFI too. The defacto per-pixel VRR behavior makes VRR essentially obsolete. The higher the Hz, the more traditional sync technologies become obsolete.

1000Hz is not the final frontier (retina refresh rate for 24" at 30-degree FOV can be well over 4,000Hz if aiming at a large-sigma of population). Accurately temporally emulating a retro display needs somewhat of an oversample margin too.

VSYNC ON, VSYNC OFF, FreeSync, G-SYNC converges into identical zero-lag zero-stutter look at all frame rates, regardless of whatever the hell the frame rate you run at. It's beautiful how throwing more Hz, sync technologies converges into identical zero-lag zero-stutter ops. Sync technologies exist only because Hz is low.

(Be noted -- there is now already a way to create 8K 1000fps 1000Hz demonstration using today's off the shelf technology -- via refresh-rate-combining multiple projectors in a round-robin fashion onto the same projector screen -- albiet requires a high budget and probably pre-rendered material initially)

IMHO, it is double plus ungood correctness to think polar like "sample and hold" vs "impulse driven" when 1000fps+ 1000Hz+ enables infinite numbers of Holy Grail BFI Modes

(e.g. CRT electron beam simulators, of a dream CRT of your choosing. Done. Or want to emulate a Pioneer Kuro plasma display instead? Done. Or 120Hz rolling OLED BFI mode? Done. Etc. Etc. Etc. Etc. Etc. The same display can emulate all retro impulsed displays, if it has retina Hz + retina HDR + retina rez)

We kind of think 10+ years out too, we're not working on what other engineers started work last year on.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: upcoming display technologies that aren't sample-and-hold?

Post by Chief Blur Buster » 22 Jun 2022, 21:42

Now that being said, I will answer an important question.

One biggest development is the boom of OLED gaming monitors in the next several years.

Several companies are about to stop manufacturing LCDs and expand the fabrication of OLEDs. This will produce a big boom of high-Hz OLED monitors in the coming years. The writing is on the wall with the launch of the Dell Alienware OLED ultrawide, but many vendors at DisplayWeek also publicly demonstrated high-Hz OLED.

Simple monolithic BFI on a 120Hz OLED or 240Hz OLED is still a pretty decent amount of display motion blur reduction for 60Hz content, so work will be advocated to strongly recommend that future high-Hz OLED always includes a BFI mode. It is the intent of a possible future Blur Busters Approved programme being made OLED-compatible, to require BFI modes to be built into future OLED gaming monitors. There are persistence limitations (e.g. BFI not being as low-persistence as LCD strobing) but the ultrafast 0ms GtG greatly compensates a lot, and 60 Hz single strobe would be a mandatory requirement.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Jefferyk2
Posts: 1
Joined: 19 Jun 2022, 08:08

Re: upcoming display technologies that aren't sample-and-hold?

Post by Jefferyk2 » 22 Jun 2022, 22:23

willrs wrote:
19 Jun 2022, 12:42
Like Mark has said, BFI is a band aid. Yes it helps but I don't know if it will ever fully eliminate motion blur and I agree it's not ergonomic (especially on OLED, I've seen BFI been a bit too flickery at times. BFI seems more ergonomic on LCD's). I think it's fair to say that if you care about motion blur, sample-and-hold displays are just inherently flawed, it's completely unrealistic to expect devices to record video at 480 or 960 frames per second, or video game consoles to render 480 or 960 frames per second. (And what about existing content?) So, my question is, is there any non-sample and hold displays being developed right now? And also, surely people involved in these decisions knew that sample-and-hold has an inherent motion blur flaw, why invest in them so much?
The ideal technology is microLED. OLED cannot get bright enough for true HDR, even after burn-in. While OLED typically has 400 to 800 nits, microLED has a maximum brightness of 10,000 nits. True HDR requires at least 1000 nits, though in some rare instances, movies are mastered at 4000 or even 10,000 nits.

A display's modular design, which appears to be the case for microLED displays, makes it simple to replace dead pixels. Simply remove and replace the dead pixel in a square area of the screen. The demonstrated microLED TVs function in a similar manner.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: upcoming display technologies that aren't sample-and-hold?

Post by Chief Blur Buster » 23 Jun 2022, 17:32

Jefferyk2 wrote:
22 Jun 2022, 22:23
The ideal technology is microLED. OLED cannot get bright enough for true HDR, even after burn-in. While OLED typically has 400 to 800 nits, microLED has a maximum brightness of 10,000 nits. True HDR requires at least 1000 nits, though in some rare instances, movies are mastered at 4000 or even 10,000 nits.

A display's modular design, which appears to be the case for microLED displays, makes it simple to replace dead pixels. Simply remove and replace the dead pixel in a square area of the screen. The demonstrated microLED TVs function in a similar manner.
Micro LED definitely has much bigger nit headroom for brightness surges, which makes retro display emulation easier (the intense brightness of a CRT electron beam).

For the software-based / videoprocessor-based / scaler-based CRT scanning emulation via brute Hz, more HDR surge headroom is nice, as the CRT beam spot size can briefly exceed 10,000 nits at the dot. Although we can spread it over time periods (e.g. 1/1000sec worth of beam simulation), the more refresh cycles available means we can use a tighter HDR window -- fewer percent of screen needs to surge HDR-maximum-ultrabright, and can thus surge brighter. At 3840fps 3840Hz using 64 refresh cycles per CRT Hz, we only need 1/64th of the screen (under 2%) to surge max-HDR-bright, as HDR will usually only go its max brightness in a tiny window size (1%-5%). If this is a 10,000nit surge for 1/64sec per pixel, it will still only average between 100-200 nits, but that's the same brightness as a typical CRT.

For designing a retro-display-simulator via sheer Hz -- one can customize the simulated phosphor fade curves to spread over multiple refresh cycles, and should, to accurately simulate decay, but the dominant persistence will be one sample-and-hold refresh cycle worth, so 4000Hz sample and hold displays will be able to do custom software/videoprocessor/scaler-based BFI algorithms down to about 0.25ms MPRT. The next refresh cycle can be a logarithmic phosphor fade (e.g. 80% or 90% decrements) -- the beauty of software-based retro display simulators is you can tweak everything (custom phosphor fade curves, or temporal dithering algorithms, or color-sequential).

Even at 120-240Hz, TestUFO has some crude algorithm simulators made possible by 2-3 refresh cycles per "final refresh cycle".

Monolithic BFI - www.testufo.com/blackframes
DLP colorwheel -- www.testufo.com/rainboweffect (do not test at less than 120Hz)
CRT Interlacing -- www.testufo.com/interlace

My new DLP colorwheel simulator works on a 360Hz LCD monitor almost exactly as well as an old Infocus DLP projector 360Hz colorwheel; with exactly the same rainbow-effect artifacts. Pretty neat how my retro-display-simulators reliably predictably become more and more accurate by throwing brute Hz at it. Brightness (lack of HDR surge headroom for extra nits) remains a big problem, though.

However, as Hz hits 1000+ and adds HDR, with more refresh cycles per simulated refresh cycle -- the number of emulatable impulse-driven displays begun to explode geometrically, with much more brightness and color-gamut match.

DMD mirrors in DLP flip at 960Hz-2880Hz, and plasma subfields often pulsed at 600Hz, etc. As more digital Hz becomes available, past displays can be more simulatable simply by brute-Hz sample and hold displays of any technology.

As the years passes, I plan to add a TestUFO temporal color dither simulator (best for 360Hz+ monitors, preferably 500Hz) and a TestUFO CRT beam simulator (best for 360Hz+ LCD or future 240Hz+ OLED), to continue TestUFO's legacy as a trailblazing demonstration platform.

Implementing a simulated 1-bit DLP temporal dithering algorithm in a GPU shader will be a very fun shadertoy exercise -- it should start producing usable results on the upcoming 500 Hz monitor; generating 24-bit color using 500Hz of 1-bit monochrome images, provides 20Hz of 24-bit color or 40Hz of 12-bit color. 500Hz is certainly not as fast as the early DLP chips but we can finally begin to do some temporal-dithering scientific-entertainment in a mere TestUFO link...

(Heck, I could even motion-compensate the temporal dithering (scroll the temporal dither at the full 500 fps=Hz, despite only being able to cycle full color at low Hz (20-40Hz)...)

Long term, brute sample-and-hold Hz will be fun for inventing your own custom "non-sample-and-hold" dream impulsed display algorithms in 100% software!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

willrs
Posts: 19
Joined: 04 Nov 2021, 18:12

Re: upcoming display technologies that aren't sample-and-hold?

Post by willrs » 24 Jun 2022, 10:57

I read someone who apparently works for a large display company say that they're not sure OLED will ever be able to do 960Hz in a consumer product, so does that mean microLED is needed for 960Hz for consumer products?

Also, how would you compare the overall performance an OLED with BFI to a Plasma? I know there are different ways of doing BFI but I'm very curious about this, I might even go buy a Plasma on ebay depending on what you say. From what I've seen very briefly on a friends plasma compared to OLED (this was not a side by side test, my friend only has a Panasonic plasma), I must say, I wasn't impressed by BFI on the OLED but maybe the type of BFI he was using was not the best?

I'm just really curious to know what is the best overall option now and in the near to medium term future, because man, when you see a 60Hz LCD next to a Plasma or a CRT, it's actually horrifying, it's unplayable.

howiec
Posts: 183
Joined: 17 Jun 2014, 15:36

Re: upcoming display technologies that aren't sample-and-hold?

Post by howiec » 11 Jul 2022, 18:06

Yeah, I was disappointed that there's no BFI feature on the new QD-OLEDs.

I have the AW3423DW and the panel is awesome compared to LCDs (I also have the AW2521H among others) but obviously the blur still exists and is significant.

Here's to hoping that the industry wakes up and realizes that there's a huge benefit to fast pixel response and refresh rates PLUS BFI resulting in superior motion clarity.

User avatar
Kamen Rider Blade
Posts: 61
Joined: 19 Feb 2021, 22:56

Re: upcoming display technologies that aren't sample-and-hold?

Post by Kamen Rider Blade » 11 Jul 2022, 22:11

I'm still rooting for SED, the true successor to CRT.

With Quantum Dots working, we could easily use Violet Phosphers as the Emission Phospher and Quantum Dots for modulating the light into RGB. Then use a Hexagonal Lattice for RGB and we have ourselves a next gen Flat Panel Successor to CRT.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: upcoming display technologies that aren't sample-and-hold?

Post by Chief Blur Buster » 11 Jul 2022, 22:25

willrs wrote:
24 Jun 2022, 10:57
I read someone who apparently works for a large display company say that they're not sure OLED will ever be able to do 960Hz in a consumer product, so does that mean microLED is needed for 960Hz for consumer products?
Never say never for OLED.

The main problem is the backplane speed, but there are many workaround and mitigations possible, if you think outside of the box.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Post Reply