Technical Limitations Preventing Ultra High Hz OLED screens

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

Technical Limitations Preventing Ultra High Hz OLED screens

Post by theTDC » 15 Mar 2021, 16:59

First off, thanks to the forum, mostly Mark Rehjon himself, for educating me in my previous two topics. My understanding of the limitations for LCD/OLED monitors, at least achieving high Hz, is as follows:

LCD:
1) Horrible GtG times. If the refresh rate is 1,000 Hz, but the GtG times can be up to 10ms, you can have frames that still have information from 10 frames ago. Garbage.
2) Scanout time.

OLED:
1) Just the scanout time. GtG times will not become limiting factors until > 1,000Hz displays.

I’m ignoring strobing technology for now. With a 1,000 Hz OLED, you could always do software blackframe insertion if for some reason you couldn’t hit 1,000 Hz with the source material, potentially multiple times in a row, at your leisure. Yes, brightness would suffer, but that’s not the focus of this post. That’s not what we’re concerned with.

My question is very simple: what are the technical limitations to just sending the scanout faster? If we had, back in 2017, an LCD monitor that could do the scanout at 480 Hz, call it 500Hz, why couldn’t we already have an OLED monitor with a 500 Hz refresh rate? On top of that, electronically, what is the actual limitation to sending the electricity to the individual pixels?

I’m very curious to see if this is a legitimately challenging technical problem, or simply something with a perceived very limited market, thus not being built for economic/demand reasons.

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by Chief Blur Buster » 15 Mar 2021, 17:22

There's no reason why OLEDs can't hit 1000 Hz eventually.

I've posted an old thread about this years ago:
Custom OLED Rolling Scans -- Custom Built OLED Monitor.

However, the first 1000 Hz gaming monitor will probably still be an LCD -- at least initially. It might be a sped-up version of an existing LCD, or one of the newer LCD technologies such as blue-phase LCDs (GtG response time in microseconds!).

P.S. I will merge this thread to the original OLED thread you've created (that I've moved into the Area51 forum area).
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: Technical Limitations Preventing Ultra High Hz OLED screens

Post by theTDC » 15 Mar 2021, 19:40

Chief Blur Buster wrote:
15 Mar 2021, 17:22
There's no reason why OLEDs can't hit 1000 Hz eventually.
I read that thread. Then I read all the spinoff threads mentioned in that thread. My original question remains. What is the physical mechanism through which the actual pixels are eventually updated? I was thinking this was something like how computer memory is updated, with there being certain bandwidth limitations. After all, if each pixel is 24 bits, and you have something like a 64 bit memory controller, you can only update 2.5 pixels per cycle. Doing that math I thought that there had to be some hardware in there that worked on this type of principle.

Now I'm not so sure, especially since you say that the entire line of pixels can be updated at the same time, and for many monitors multiple lines are updated simultaneously. Now I definitely don't understand how this works, but I'm wondering if there is some limit? If we can update 2 lines simultaneously, why not 10?

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by MCLV » 16 Mar 2021, 03:24

There is no reason why many lines (or even all of them) cannot be updated simultaneously. However, you have to understand that consumer products are not focused purely on overcoming technical limitations. That is done in research labs. In commercial world, it needs to be technically feasible and cheap. Huge emphasis is on the cheap part. It also often important that the technical solution is "proven". I put quotation marks there because it can mean that it is proven to work just passably. I sometimes say that these methods are proven in a way that it's proven that it doesn't really work as it should :) But any better alternative means additional development cost, risk and maybe even 1$ increase of bill of materials/parts. And you better find a really good argument to explain this to management :lol: Whether you can do it depends a lot on the industry, company culture, personality of your superiors and also on your ability to explain and communicate.

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by MCLV » 16 Mar 2021, 05:54

By the way, one practical limitation to refresh rate comes from display interface. To keep cost down, you want to use existing standards. 1920x1080 @ 480 Hz results in 28 Gbps which became possible only recently with HDMI 2.1 (HDMI 2.0 or DP 1.4 are not enough). 1920x1080 @ 960 Hz would need 56 Gbps of bandwidth and there is no interface that can handle it until DP 2.0 arrives. And it would also need UHBR 20 transmission mode which I'm not sure will be supported by first generation DP 2.0 products (similarly like not all HDMI 2.1 devices support full data rate today). And the situation is obviously much more limiting at higher resolutions and also with 10 bits per channel to support HDR.

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by theTDC » 16 Mar 2021, 14:44

MCLV wrote:
16 Mar 2021, 03:24
There is no reason why many lines (or even all of them) cannot be updated simultaneously. However, you have to understand that consumer products are not focused purely on overcoming technical limitations. That is done in research labs. In commercial world, it needs to be technically feasible and cheap. Huge emphasis is on the cheap part. It also often important that the technical solution is "proven". I put quotation marks there because it can mean that it is proven to work just passably. I sometimes say that these methods are proven in a way that it's proven that it doesn't really work as it should :) But any better alternative means additional development cost, risk and maybe even 1$ increase of bill of materials/parts. And you better find a really good argument to explain this to management :lol: Whether you can do it depends a lot on the industry, company culture, personality of your superiors and also on your ability to explain and communicate.
All of them? Through what physical process does this happen? Even when writing data to memory a CPU can’t just write to all of them at the same time, so how does that work? We would need a direct line linking the memory the image is stored in with the pixel itself. That seems... like a whole lot of wires for a 1080p display. About 2.075 million of them, directly hooked up to the memory.

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by MCLV » 16 Mar 2021, 16:43

Maybe I worded it badly. And I would also like to point out that I don't have deep knowledge of panel electronics so I might be wrong.

What I meant is that I imagine that in the simplest implementation, you select one row of pixels and then you go through all columns in that row and set pixels one by one. To speed it up, you could update on multiple rows of pixels at the same time (which I understood can be done at the moment). So I thought that if can do 2 lines at once you could in principle scale it up and do 4, 8 and eventually all of them in parallel. However, even then you wouldn't update all pixels at once since still have to go through all columns separately.

And even the fact that it could be done doesn't mean that it would be very practical thing to do. Because the resulting device would be very likely too expensive, power hungry, needing active cooling and also quite bulky. If all that doesn't stop you would be able to get what you want.

By an analogy, if you would ask if there are any technical limitations limiting LHC at CERN to be "only" 27 km in circumference. Why not 50 or 100 km? I would answer you that there is nothing in principle preventing us to build a 100 km collider. Well, apart from persuading taxpayers to fund it.

In the end, I'm not sure if your question is if 1000 Hz OLED screen is possible at any price point or why you cannot buy it for $500 for example...

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

Re: Technical Limitations Preventing Ultra High Hz OLED screens

Post by Chief Blur Buster » 16 Mar 2021, 16:57

theTDC wrote:
16 Mar 2021, 14:44
MCLV wrote:
16 Mar 2021, 03:24
There is no reason why many lines (or even all of them) cannot be updated simultaneously. However, you have to understand that consumer products are not focused purely on overcoming technical limitations. That is done in research labs. In commercial world, it needs to be technically feasible and cheap. Huge emphasis is on the cheap part. It also often important that the technical solution is "proven". I put quotation marks there because it can mean that it is proven to work just passably. I sometimes say that these methods are proven in a way that it's proven that it doesn't really work as it should :) But any better alternative means additional development cost, risk and maybe even 1$ increase of bill of materials/parts. And you better find a really good argument to explain this to management :lol: Whether you can do it depends a lot on the industry, company culture, personality of your superiors and also on your ability to explain and communicate.
All of them? Through what physical process does this happen? Even when writing data to memory a CPU can’t just write to all of them at the same time, so how does that work? We would need a direct line linking the memory the image is stored in with the pixel itself. That seems... like a whole lot of wires for a 1080p display. About 2.075 million of them, directly hooked up to the memory.
To both of you --

RAM and screens are refreshed in similar ways -- row-and-column addressing technique.

This reduces the number of kilometers of microwires needed, since you only need to light up one vertical wire and one horizontal wire -- the intersecting area is the subpixel being controlled.

But screens are special in a way in that they need some more extreme optimization needed due to the long distances of those panel microwires carrying fairly high voltages (>10 volt) to remote transistor gates that need to be run long enough to cause the active transistor to switch. But the less time you have per pixel, the less power per active matrix transistor and GtG becomes slower at higher refresh rates, unless you do clever workarounds.

Techniques have been done to increase the number of channels of simultaneous pixel refreshing, with dramatic increases in panel electronics complexity.

Can't Refresh All Pixels At Same Time

Practically, there's definitely no way to refresh all pixels simultaneously. But we don't need to. We just need to give a good fast voltage kick to the pixels at line refresh. GtG is physical momentum of LCD molecules of a Liquid Crystal Display -- rotating molecules behaving as light valves to block/unblock polarized light.

See LCD GtG as Soccer Balls Metaphor to understand the law-of-physics behavior needed for fast GtG.

Each LCD pixel needs a running start for a hard fast GtG kick, so the soccer ball can fly fast on its own momentum while you're starting ar running start for the NEXT soccer ball (next LCD pixel).

The LCD pixel soccer ball kick lasts less than a microsecond!

If we tried to kick 2 million soccer balls simultaneously with strong momentum, there will be massive side effects such as voltage droop from the big instantaneous power surge over millions of microwires, crosstalk inteference between adjacent voltage surges, not to mention the (multi)million of dollars (per panel) worth of concurrent chips/processors necessary to do 2 million pixels realtime simultaneously with thousands of kilometers worth of microwires wired concurrently to multiple chips running in parallel to operate 2 million soccer ball kicks concurrently. A gaming monitor for well over $100M prototype or $1M mass-manufactured. Want one?

Now, if we could refresh all pixels simultaneously in just 1 microsecond, the chips then idle -- for 999999/1000000ths of a second. What's stopping it from being able to refresh at 1 million hertz? (Blue-phase LCDs can have GtGs in microseconds). Inefficient utilization of chip resources. Waste, waste. Why?

But There's Another Reason: Worse Input Lag If We Refresh Panel Globally

Also, it's more efficient latency-wise to stream the cable directly onto the panel. Since 1920s analog televisions, we've been serializing 2D data into 1D, via a raster mechanism. All modern signals are serializations of 2D data into 1D -- Blur Busters Custom Resolution (CRU) Glossary 101 / FAQ ...

I can even get FreeSync working on an analog CRT because it's just essentially a variable-thickness VHOLD black bar spacering between refresh cycle. It's impressive how minor a signal modification raster VRR is, and how much in common a 2020's DisplayPort signal has with a signal being used for Baird 1920s prototype Nipikow mechanical TV.

For ~100 years, we've been sticking to the same signal methodology of a serialization of 2D-into-1D (wire/broadcast/analog/digital/delivery/file/whatever). Heck, we've been rasterizing for more than 150 years if we don't neglect to consider those 1860s-1870s mechanical fax machines (pendulum-powered pantelegraphs and resulting raster images), which is also a raster-based 2D delivery of image data over a 1D telegraph wire. Anyway, back to today.

Most gaming LCDs have optimized latency by syncing panel scanout to cable scanout. VGA/HDMI/DVI/DP are identical signal-topologically as a calendar-readout or book-readout top-left through bottom-right, including sync intervals (original CRT beam movement signals, but now behaves as digital indicators and guard delays). Even Baird/Farnsworth 1920s TV had sync intervals too, and they still exist today in 2020s DisplayPort.

Image

Image

Image

So we'd have MORE input lag if we refreshed all pixels at the same time, because we have to buffer the slow-delivering signal before we had all the image data to refresh the whole screen instantly with. We can't transmit data infinitely fast over the cable.

And lest, skeptical scientists hand-wave-off the importance of the millisecond, it is important they read The Amazing Human Visible Benefits Of The Millisecond to understand the implications better.

TL;DR: Practically, it just isn't economically possible to refresh all pixels simultaneously
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: Technical Limitations Preventing Ultra High Hz OLED screens

Post by Chief Blur Buster » 16 Mar 2021, 17:01

Whoops, our messages crossed. An edit of my post just posted after your post.
MCLV wrote:
16 Mar 2021, 16:43
What I meant is that I imagine that in the simplest implementation, you select one row of pixels and then you go through all columns in that row and set pixels one by one. To speed it up, you could update on multiple rows of pixels at the same time (which I understood can be done at the moment). So I thought that if can do 2 lines at once you could in principle scale it up and do 4, 8 and eventually all of them in parallel. However, even then you wouldn't update all pixels at once since still have to go through all columns separately.
That's already being done in many panels -- first ones were dual-scan which refreshed the top/bottom halves concurrently. This also made the first 4K LCD possible (IBM T-221) -- great teardown video, where FPGAs attached to ribbon cables connecting top and bottom to allow concurrently refreshing different parts of the panel.

Since then, panels have gained more channels like interleaved 2 or 4 channel refresh -- where 4 contiguous pixel rows are refreshed concurrently (4 pixels at a time). Think of it like 4 independent sets of row-column addressors, like 4 separate RAM chips, except 4-way-interleaved into each other like [1A,2A,3A,4A],[1B,2B,3B,4B],[1C,2C,3C,4C] denoting the first 12 pixel rows of an LCD. 1A,1B,1C,... goes to one row-column addressor, 2A,2B,2C,... goes to a different row-column addressor, and so on. So modern digitally-addressed are already doing these sorts of optiimzations for LCDs, OLEDs, MiniLEDs, MicroLEDs, and JumboTrons (in various interleave patterns appropriate to their refreshing workloads).

This also helped with addressing-modification tricks like 480 Hz overclocking of a 4K 120 Hz panel by binning 4 pixel rows into quadruple-speed refreshing. That 4K 120Hz panel was hacked to do 1080p at 240Hz, and 540p at 480Hz. Basically the electronics was refreshing 4 pixel rows simultaneously with the same pixel data, for the same amount of time it would normally refresh 1 pixel row (i.e. 4 separate pixel rows with different pixel data). This was a real hack to do such as massive overclock of the LCD via abuse of the row-column latching to force refresh electronics to refresh 4-in-1 in the time interval of refreshing 1, with the only side effect of being a vertical quartering of resolution.

It's still top-to-bottom scan (best optimization to sync to cable delivery), even though it is chunking multiple-pixel-row. This is not visible in high speed videos because LCD GtG is too slow to reveal this, especially since it'll have refreshed (initiated the "GtG kick") over a hundred pixel rows ahead in the time of LCD GtG coasting on its own sheer momentum.

Some panels, especially LED matrixes, even have full-row-refresh through some creative tricks (shift registers, etc) but it's still contending to the law-of-physics of 1D delivery (albiet very rapid). Lots of tricks layered on top of each other. Even "global refresh" displays like plasmas and DLP are simply ultra high velocity low-precision scanouts (e.g. low-bit-precision temporal dithers, like DLP 1/1440sec side-to-side raster sweep of 1-bit monochrome image in a single DLP "field"). They have way more lag than a gaming LCD which have heavily optimized to syncing 1:1 between cable scanout and panel scanout.

I expect channelized refresh tricks to increase -- 8 channel, 16 channel -- to help the refresh rate race -- though pressures of thinner bezels make this hard -- given many circuits are embedded into the panel on the edges. An LCD is simply one giant integrated circuit which can contain some processing (essentially on-glass chips) to build things like shift registers, though number of layers can be limited (wire-over-wire, etc), which can limit the number of interleaved channels achievable.
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: Technical Limitations Preventing Ultra High Hz OLED screens

Post by theTDC » 16 Mar 2021, 17:55

Chief Blur Buster wrote:
16 Mar 2021, 17:01
This also helped with addressing-modification tricks like 480 Hz overclocking of a 4K 120 Hz panel by binning 4 pixel rows into quadruple-speed refreshing. That 4K 120Hz panel was hacked to do 1080p at 240Hz, and 540p at 480Hz. Basically the electronics was refreshing 4 pixel rows simultaneously with the same pixel data, for the same amount of time it would normally refresh 1 pixel row (i.e. 4 separate pixel rows with different pixel data). This was a real hack to do such as massive overclock of the LCD via abuse of the row-column latching to force refresh electronics to refresh 4-in-1 in the time interval of refreshing 1, with the only side effect of being a vertical quartering of resolution.

...

I expect channelized refresh tricks to increase -- 8 channel, 16 channel -- to help the refresh rate race -- though pressures of thinner bezels make this hard -- given many circuits are embedded into the panel on the edges. An LCD is simply one giant integrated circuit which can contain some processing (essentially on-glass chips) to build things like shift registers, though number of layers can be limited (wire-over-wire, etc), which can limit the number of interleaved channels achievable.
So basically if I'm understanding this correctly, the physics of sending the correct voltage down the wires is such that we can only do one pixel at a time for each "channel" we have, with some minor exceptions. However, the number of channels that we can have are theoretically arbitrary as long as we are fine with extremely wide bezels. Maybe some other (minor?) issues such as increased voltage/power consumption per second in setting the pixels.

Currently the scanout speed appears to be limited to a maximum of ~8ms for an entire 4k display when only using a single channel. Assuming this scales linearly, 2 channels would give us 240Hz, 4 channels 480Hz, and 16 channels could give us potentially 1,920 Hz. If that's all true, then what's holding back Ultra High Hz monitors, at least as a niche product, is not the scanout speed. For LCDs it's the GtG speed, and for OLEDs and LCDs it's the monitor cable, plus the arguable lack of content to justify such high Hz displays.

Post Reply