Why Can't/Don't OLED's have short light times?

Advanced display talk, display hackers, advanced game programmers, scientists, display researchers, display manufacturers, vision researchers & Advanced Display Articles on Blur Busters. The masters on Blur Busters.
theTDC
Posts: 25
Joined: 09 Mar 2021, 00:13

Why Can't/Don't OLED's have short light times?

Post by theTDC » 09 Mar 2021, 23:34

I've looked all over this website, but I just can't seem to find the drawbacks of OLEDs. I understand that many of them seem to just hold the frame for the entire duration of the frame, which will lead to massive sample/hold blur. However, Oculus and a few others have OLEDs that have a way of quickly shutting off the pixels after they get lit up. So my questions are the following:

1) Is there some technical difficulty with having a faster scanline/progression on an OLED? If we have 100Hz OLEDs, and ~500 Hz LCDs, why can't we update the OLED pixels at similar rates easily?

2) Is there anything stopping us from going even lower. Basically just send a second signal down the monitor 100 microseconds after the first (or whatever time we want) that shuts the pixel off?

3) OLEDs suffer when forced to go too bright. If we have one that spends 1ms on at 4x luminence, versus 4ms on at 1x luminence, for the same brightness is the former harder on the pixels than the latter?

I'm just struggling to understand why, other than potentially financial/business reasons, we don't just all have ultra low persistence OLED displays. I'm thinking maybe it's just a brightness thing, where LCDs can go brighter, so they suffer less from strobelights.

iopq
Posts: 48
Joined: 23 May 2020, 10:06

Re: Why Can't/Don't OLED's have short light times?

Post by iopq » 11 Mar 2021, 07:37

Some OLEDs can't turn on as fast, so they have dark smear because when the pixels are completely off they take too long to come back on. Of course these days it doesn't seem to be as much of an issue, as BFI is offered with OLED TVs

I just want my high refresh OLED monitor!

theTDC
Posts: 25
Joined: 09 Mar 2021, 00:13

Re: Why Can't/Don't OLED's have short light times?

Post by theTDC » 12 Mar 2021, 14:31

So, if I’m understanding you correctly, historically there were some good reasons because the OLED pixels couldn’t actually switch that fast if they were going from completely dark? Okay, but that appears to be solved, so now we’re back to square one.

I guess maybe I should have started my question a little differently. Aside purely from financial/business reasons, what are even the drawbacks of OLED screens? I’ve looked here and on Wikipedia, and I can’t seem to find any good reasons. And on this site it would appear to be even more of a no-brainer, since the fast pixel switching times would allow you to easily have low persistence displays with high refresh rates.

It appears that Oculus does this, so why doesn’t anyone else? Why doesn’t everyone else?

MCLV
Posts: 43
Joined: 04 Mar 2021, 15:04

Re: Why Can't/Don't OLED's have short light times?

Post by MCLV » 13 Mar 2021, 06:44

You can find disadvantages of OLED screens here: https://en.wikipedia.org/wiki/OLED#Disadvantages

Lifespan issue has pretty big implications and I think that it answers basically all your questions. Higher the brightness, the shorter the lifespan is. And the wear of pixels likely increases faster than their brightness. Otherwise, manufacturers could easily use strobing/PWM to achieve higher brightness and longer lifetime. But they don't. Hence, there is some practical brightness limit so you can't strobe with ~1 ms pulse duration at low refresh rates and achieve sufficient brightness. However, black-frame insertion seems feasible.

Lifespan issues also mean that OLED PC monitors are not widespread. Typical PC usage has a large amount of static content and also >8 hours a day if you use it for work as well. This results in potentially very high risk of burn in. OLED displays are getting better and better in this regard but it seems that manufacturers are still not comfortable to introduce them for mainstream PC market.

This means that pressure to introduce high refresh OLED displays is low since TV and movie industry is stuck at low refresh rates. They could be used for gaming but console gaming is also not exactly known for pushing high refresh rate. In my opinion, unless OLED displays become usable for general PC market and PC gaming, there is little incentive for manufacturers to introduce high refresh (e.g. 240 Hz or more) OLED displays. At least when were are talking about large screen sizes and not about mobile or VR devices.

theTDC
Posts: 25
Joined: 09 Mar 2021, 00:13

Re: Why Can't/Don't OLED's have short light times?

Post by theTDC » 13 Mar 2021, 15:19

MCLV wrote:
13 Mar 2021, 06:44
You can find disadvantages of OLED screens here: https://en.wikipedia.org/wiki/OLED#Disadvantages
Well, not to be a dick, but I did mention that I had already looked through the wikipedia article before. I was already aware of those problems, and wanted to see if someone actually knew for sure if the OLED burn in times were exponential related to their brightness, versus simply linear.

Also, why Black Frame Insertion? Why can't we just quickly send a signal to turn the pixels off after some arbitrary and short period of time?

I am aware that business reasons are probably the main reason why we don't all have super high Hz OLEDs. There is something to be said for a bright, cheap LCD in an office situation where it's just being used to display information. After all, that's why LCDs were invented. I wanted to know if there were other more serious technical limitations.

MCLV
Posts: 43
Joined: 04 Mar 2021, 15:04

Re: Why Can't/Don't OLED's have short light times?

Post by MCLV » 13 Mar 2021, 17:02

theTDC wrote:
13 Mar 2021, 15:19
Why can't we just quickly send a signal to turn the pixels off after some arbitrary and short period of time?
I think that black frame insertion is implemented in this way in OLED panels. However, the duty cycle is fixed 50% ON and 50% OFF. In my opinion, it should be technically possible to have shorter ON period but this would likely result in unacceptably low brightness and this is the reason why it is not done like that.
theTDC wrote:
13 Mar 2021, 15:19
if someone actually knew for sure if the OLED burn in times were exponential related to their brightness, versus simply linear.
Regarding the lifetime and brightness relation, I have found two open access papers:
1) "Thermal behavior and indirect life test of large-area OLED lighting panels" by Huiqing Pang, Lech Michalski, Michael S Weaver, Ruiqing Ma and Julie J Brown https://core.ac.uk/download/pdf/81834393.pdf
Authors claim that wear increases with luminance to power 1.43.

2) "Parametric lifespan models for OLEDs using Design of Experiments (DoE)" by Farah Salameh, Antoine Picot, Laurent Canale, Georges Zissis, Marie Chabert, Pascal Maussion https://hal.archives-ouvertes.fr/hal-02319746/document
Where authors claim higher than linear wear increase in both current density and temperature.

These papers do not claim exponential relationship but both claim that it is superlinear. Which means that at equal average brightness, short pulses with high brightness will result in higher wear than always ON with lower brightness.

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

Re: Why Can't/Don't OLED's have short light times?

Post by Chief Blur Buster » 13 Mar 2021, 17:29

MCLV wrote:
13 Mar 2021, 17:02
I think that black frame insertion is implemented in this way in OLED panels. However, the duty cycle is fixed 50% ON and 50% OFF. In my opinion, it should be technically possible to have shorter ON period but this would likely result in unacceptably low brightness and this is the reason why it is not done like that.
The Talbot-Plateau Law is the biggest problem though. You need to flash twice as fast to keep brightness. 0.1ms strobe requires 167x the brightness of 16.7ms 60Hz OLED, to keep the same brightness as a 60Hz display. That can burn out an OLED quickly. But LED lighting has already become super bright, they light up stadiums now. It's now possible to put a water-cooled or fan-cooled LED backlight behind an LCD, so the strobing scales better with LCDs than OLEDs.

Also, S/N ratio increases for more brightly strobed OLEDs. OLED dark-screen banding/streaking also becomes worse if you try to brighten the OLED image to compensate for the short flash times. Thus, you have smaller dynamic range between noisy blacks and the brightest whites -- and/or more grainy/noisy image since your strobe:GtG ratio becomes too tight for good OLED strobing.

Remember, OLED strobing means strobes are directly OLED GtG powered, and OLED GtG isn't infinitely fast. 0.1ms GtG is still 10% of a 1ms MPRT strobe. Any GtG non-uniformities (e.g. greens being 0.2ms GtG and reds being 0.13ms GtG) creates color distortions during brief strobes. LCD strobes are backlight powered (independent of the LCD panel), and discrete LEDs are very fast at switching (especially RGB LEDs that can switch millions times per second -- since LEDs are used to power some optic fibers). So the outsourcing of LCD persistence tricks only require LCD GtG to be sufficiently fast between refresh cycles, so that LCD can be strobed. There's no real limit to how clear an LCD can be -- it's completely unbounded -- 1ms, 0.1ms, 0.01ms -- it's simply limited by the brightness of the strobe flash, once you manage to hide GtG between refresh cycles, GtG ceases to be a physics limiting factor for motion blur when completely hidden in the dark periods, as a synchronized co-operation between LCD layer and the backlight strobe.

This is why LCD VR (0.3ms MPRT) now currently has much less motion blur than OLED VR (2ms MPRT). It's quite clear, as I have both an Oculus Rift VR headset (2ms MPRT strobed OLED) and a Oculus Quest VR headset (0.3ms MPRT strobed LCD). And LG OLED can't get blur as low as the original Oculus Rift CV1 headset.

Further improvements to OLED will happen, but it gets very difficult with the laws of physics (Talbot-Plateau Law) and OLED limited brightness. You need 16,667 nits to keep 0.1ms 60 Hz strobe 100 nits brightness, and you need 8,333 nits to keep 0.1ms 120Hz strobe at 100 nits brightness.

I've seen the Sony 10,000nit prototype LCD, and no OLED can get as bright as that.

CRT electron beam dot beam can briefly exceed 30,000nits on your everyday run-of-the-mill 28" RCA TV or 25" Zenith TV, though it decays quickly after that (phosphor decay) -- but it is a great demonstration of extreme brightness needed to lower persistence to sub-millisecond levels.
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!

theTDC
Posts: 25
Joined: 09 Mar 2021, 00:13

Re: Why Can't/Don't OLED's have short light times?

Post by theTDC » 14 Mar 2021, 15:33

Great answers from both of you. Nice to know that my suspicious of exponential burn in from increased luminance were confirmed.
Chief Blur Buster wrote:
13 Mar 2021, 17:29
This is why LCD VR (0.3ms MPRT) now currently has much less motion blur than OLED VR (2ms MPRT). It's quite clear, as I have both an Oculus Rift VR headset (2ms MPRT strobed OLED) and a Oculus Quest VR headset (0.3ms MPRT strobed LCD). And LG OLED can't get blur as low as the original Oculus Rift CV1 headset.
So if I'm understanding this correctly, LCD technology can have almost arbitrarily bright LED backlights, letting us go pretty much as low for persistence as we could reasonably want. At least the state of the art LCDs do. However, we can't just give that to OLED displays, because then they'll burn out and die. This is a shame, because GtG times harshly limit our LCD refresh rates, but not their persistence times.

So currently we already have LCD displays with ~0.3ms persistence, corresponding to ~3,000+ Hz, and with screen refresh rates of close to 500 Hz, although the latter is misleading. Now, the oculus displays have 120 Hz refresh rates, but I don't think there's anything stopping us from going higher, other than the pointlessness for VR, since it is deemed too difficult to reach those high frame rates. Yes, ignoring FRAT and other ways of getting there for now.

OLEDs on the other hand just flat out can't be too bright. But what I don't understand is why we can't have extremely high refresh rates on them. I understand the brightness limitations, but is there any technological limitation for the refresh rate itself? If we have an OLED with no dimming, that's constantly on for, say, the entire 2m duration of a 500 Hz refresh rate, the brightness only needs to be the same as if it had a 100 Hz refresh rate. You would know better than I what the experience would be like of a ~500Hz display with full persistence.

To be clear, what I'm asking is if the scan-out technology is somehow different for the OLEDs, or if it's just the same for all.

My understanding now is as follows:

1) OLED. Limited in refresh rate by ability to scan out the pixels top to bottom and change them. Actual GtG times not close to the limiting factor. Low persistence at much less than half not practical.
2) LCD. Also somewhat limited by scan speed. However, at least 10x more limited by GtG times. Arbitrary low persistence (within reason) easily possible because strong backlight not damaging.

The fastest consumer displays are 480 Hz, so I'll say that 2ms is currently the maximum scan out speed for a normal resolution display. With that in mind, we could just make an OLED screen with a refresh rate of ~500Hz, since the 0.1-0.5 pixel colour switching speeds is not that big of a deal. For LCDs, we can also only make a ~500 Hz display, but we might want to go much lower than that, since the best worst-case pixel GtG times are going to be ~5ms. We unfortunatly have to add the scan-out time to that, so the fastest we can have a frame is roughly 7ms or so. So we're limited to something like ~150 Hz at the absolute maximum. However, in exchange, we can have persistence theoretically as low as ~0.1ms with existing technology.

I completely understand why you say that real world tests show that 120Hz LCD ultra low persistence is superior to 120 Hz OLED with ~2ms persistence. But do you think that a theoretical ~500Hz full persistence OLED would feel superior to a ~100 Hz ultra low persistence LCD? Curious to hear your thoughts.

Finally, what is the realistic best GtG times we can get for our LCDs? Is 1ms possible within a few years? What's the best rate currently?

theTDC
Posts: 25
Joined: 09 Mar 2021, 00:13

Re: Why Can't/Don't OLED's have short light times?

Post by theTDC » 14 Mar 2021, 15:48

OLED.
1) Can have higher refresh rates, since the only limiting factor is scan out. If scan out technology gets to 1ms, you can have 1,000 Hz full persistence OLEDs. GtG worst case times should be less than 0.5ms, so it's still not a hugely limiting factor.
2) However, you will always have the bottom of the screen be delayed behind the top, so things that are in motion will appear to somewhat warp, where the top is moving a bit ahead of the bottom. With a 1,000 Hz display this will probably not be that bad. With a 100Hz display, this could be very bad.
3) Low persistence isn't nearly as good. It enhances the low-brightness problem. On top of this, we can't hide pixel transitions like we can with an LCD. (This might be wrong, but then we need a second scanout beam ~0.5 seconds behind)

LCD.
1) Horrible GtG switch times. Easily an order of magnitude worse than OLEDs.
2) However, massively brighter lighting from backlights allows us to have tiny persistence, an order of magnitude better than OLEDs.
3) Pixel transitions, while far longer, can be completely hidden, and we never have the "jello" effect of an OLED, since we flash the entire screen at the same time.

Pretty fascinating stuff, and answers my original question. If low persistence is what matters, and the source material is at low framerates, an LCD is definitely where it's at. If high refresh rates matter, but high persistence is okay, and cost is no issue, then OLEDs are the answer.

blurfreeCRTGimp
Posts: 42
Joined: 28 May 2020, 20:36

Re: Why Can't/Don't OLED's have short light times?

Post by blurfreeCRTGimp » 15 Mar 2021, 19:13

"I've seen the Sony 10,000nit prototype LCD, and no OLED can get as bright as that."

Chief, I had a thought. That 10,000 nit backlight has to be power hungry, and hot.

Would it be feasible to take a MEMS laser scanning system and instead of drawing an image directly (like a Nebra Anybeam) shine the lasers through some optics onto a white phosphor layer to get really bright light at low wattage? (kind of replicating BMW laser headlights, but combined with MEMS?)

It would mimic the raster scanning of a CRT, (so contrast would be great wherever the beam was off) and with 10 watts of lasers you could probably get 10,000+ nits on a beam easily.

Feasible? Crazy? What do you think?

Post Reply