Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

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!
User avatar
saw141
Posts: 78
Joined: 16 Aug 2018, 17:22

Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by saw141 » 21 Oct 2018, 13:26

Touch Taiwan was in late August, AUO apparently showed off their upcoming 0.5ms 240hz panels. However, I’m having trouble finding information about them or any more details. Does anyone know any more news on these? They’re projected to begin development in Q4 2018, so now. Hopefully we see them soon, I’m very excited.

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

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by Chief Blur Buster » 21 Oct 2018, 15:01

These AUO 240Hz 0.5ms panels are also eagerly awaited by Blur Busters.

Many websites claim that 1ms GtG doesn't matter. 1ms and 2ms GtG may not matter at 60Hz since 60Hz = 1/60sec = 16.7ms .... but 1ms GtG certainly become a bigger chasm of a problem at 240Hz, 480Hz and up. As those smart people like us here at Blur Busters already know... fast GtG becomes more important again at 240Hz+ as 1ms GtG starts enroaching half a refresh cycle at 480Hz -- pixel response limitations starts rearing their ugly head when GtG90% starts approaching (And exceeding) 50% of the duration of a refresh cycle. At 480Hz, refresh cycles are only 2ms each, and that's where 1ms GtG90% starts showing major limitations.

While these will be used for 240Hz at first -- I am hoping some vendor (or indie maker creativity) starts overclocking these to 480Hz full-1080p by 2020. The current 240Hz panels do 240Hz much better than some yesterday 120Hz panels doing 144Hz, and my tests on an experimental 480Hz-overclocked 4K LCD -- for an LCD panel normally designed for 4K 60 Hertz monitors. A custom TCON can do wonders. 0.5ms response is more than fast enough to produce fully usable high quality 480Hz within TN LCD abilities.

This will produce better quality 240Hz. However, don't expect 480 Hz 1080p on these panels for at least one or two more years. The custom TCON necessary to do 480Hz 1080p is still expensive -- possibly more costly than GSYNC -- so expect another (approx) couple years of 240Hz.
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
saw141
Posts: 78
Joined: 16 Aug 2018, 17:22

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by saw141 » 21 Oct 2018, 21:21

Chief Blur Buster wrote:These AUO 240Hz 0.5ms panels are also eagerly awaited by Blur Busters.

Many websites claim that 1ms GtG doesn't matter. 1ms and 2ms GtG may not matter at 60Hz since 60Hz = 1/60sec = 16.7ms .... but 1ms GtG certainly become a bigger chasm of a problem at 240Hz, 480Hz and up. As those smart people like us here at Blur Busters already know... fast GtG becomes more important again at 240Hz+ as 1ms GtG starts enroaching half a refresh cycle at 480Hz -- pixel response limitations starts rearing their ugly head when GtG90% starts approaching (And exceeding) 50% of the duration of a refresh cycle. At 480Hz, refresh cycles are only 2ms each, and that's where 1ms GtG90% starts showing major limitations.

While these will be used for 240Hz at first -- I am hoping some vendor (or indie maker creativity) starts overclocking these to 480Hz full-1080p by 2020. The current 240Hz panels do 240Hz much better than some yesterday 120Hz panels doing 144Hz, and my tests on an experimental 480Hz-overclocked 4K LCD -- for an LCD panel normally designed for 4K 60 Hertz monitors. A custom TCON can do wonders. 0.5ms response is more than fast enough to produce fully usable high quality 480Hz within TN LCD abilities.

This will produce better quality 240Hz. However, don't expect 480 Hz 1080p on these panels for at least one or two more years. The custom TCON necessary to do 480Hz 1080p is still expensive -- possibly more costly than GSYNC -- so expect another (approx) couple years of 240Hz.


I've seen your article about the 480hz but I greatly appreciate you linking it and discussing it. I'm incredibly excited for 480hz and I dream of the day when we have them, how amazing. I'm also excited by upcoming Micro-LED technology, it's certainly getting even more and more interesting out there.

zivko
Posts: 16
Joined: 19 Sep 2018, 04:44

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by zivko » 22 Oct 2018, 12:35

http://www.tftcentral.co.uk/blog/asus-v ... /#more-190
apparently Asus VG258QR is supposed to use new 0.5ms g2g AU Optronics panel but this monitor only has 165hz refresh rate. considering AUO roadmap lists 0.5 ms screen as 240hz why is this only 165hz? does this means if this is the same panel as in msi nxg251 then it wont be possible to oc it past 240hz or at least not by much?

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

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by Chief Blur Buster » 22 Oct 2018, 14:13

Please remember that the 4K60 panel used in the Zisworks 4K120 is a 60Hz 4K panel (with effectively a 120Hz "overclock" capability) that successfully runs at 240Hz 1080p and 480Hz 540p with a custom motherboard (with FPGA) driving the panel.

You can view this as either a 4:1 or 8:1 overclock margin, thanks to custom TCON motherboards driving what is otherwise an inexpensive 4K panel ($80 factory cost for the panelglass before attachment of expensive custom FPGA-driven TCON)

165Hz does not mean the panel can't do 480Hz at native resolution.

There are many monitors sharing the same 24 inch 1ms TN panel for 60Hz, 144Hz and 240Hz. For a few years now, "24 inch 1ms GtG TN" panels were essentially the same panel, just with different electronics/circuitboard (TCON) mounted into the back surface of the LCD glass.

The panel glass and LCD pixels were the same, except different TCON (Timing Controller) motherboards attached to the back of the LCD panel glass.

Occasionally, lots of headroom is accidentally revealed by unexpectedly overbuilt or overclock-resistant electronics like the Successful 60Hz -> 180Hz overclock of a laptop LCD. But usually, most electronics attached to the LCD glass is the limiting factor. The panelglass & LCD pixels are usually NOT the limiting factor -- at least when it comes to TN-technology LCDs.

There is a lot of headroom in a TN LCD, the difficulty is making sure that the panel is designed to function with getting enough voltage to the pixels quickly enough, during ultrafast-refresh-cycle passes, from a custom motherboard beyond reference design. Getting a 0.5ms GtG curve permits the ability to operate a rapid-scanout-pass workflow -- hopefully without generating too many coronas/overdrive artifacts.

LCD panels are not like CPUs. They "don't" have a native refresh frequency -- that's the motherboard's responsibility. Better GtG capabilities for the same pixel-volting duration (Preferably GtG90% metric for all 256x256 color combos) means it's easier for custom electronics to drive the LCD faster (briefer volting per pixel row during high speed scanout) for various advantages -- such as high velocity scanout for long VBIs for clearer strobe -- and/or at higher frequency (aka more refresh cycles).

The faster/briefer you try to refresh an LCD, the more they smear. That's what happened when LCDs were so slow (33ms) twenty years ago, that's more than a 60Hz refresh cycle. In other words, metaphorically (although we never called it that), we were essentially "overclocking" the LCDs to get just 60 Hz. The smearing behaviour begins when refresh cycle durations starts enroaching roughly half GtG.

The threshold varies a bit, as there are other laws-of-physics at play (e.g. resistance of the microwire grid in the LCD panel). Custom motherboards can increase pixel refresh voltage to compensate partially for the briefer pixel refresh, and many other tricks -- then you start hitting a brick wall where you can't jack up voltage per pixel any further, nor can refresh a pixel any briefer. However, these brick wall is not often hit from a custom TCON/motherboard-perspective until a many-times-to-one ratio, for well-engineered TN LCD panel glass.

Also, binning is done sometimes. Not always. Panel validation is done, testing 144Hz or 240Hz panels to see if they show artifacts, and if they do, they are rebinned for lower refresh rates (e.g. 60Hz) and driven by cheaper TCONs. Unlike CPUs, the Hertz gradient is huge -- can be practically one-order magnitude. This does not always happen, since some well made 60Hz 1ms TN panels are reliable at 144Hz when doing the TCON switcheroo, e.g. transplanting between 60Hz and 144Hz monitors using the same panel (takes a lot of work, because it's hard to rip off a TCON from an LCD panel). This playing around often costs 2x the price of a cheap 144Hz monitor, so enthusiasts don't do this kinda stuff. Just the display hardware hackers like Zis. Binning may need to occur for 480Hz though, but it's usually no longer necessary for 1ms TN panels, they are just indiscriminately sold from the panel factory for 60Hz/144Hz without needing Hz validation pretests for most manufacturers now. Panel glass without TCON is relatively cheap, costing only few tens of dollars for your garden variety 24" TN 1ms LCD. (The 0.5ms is probably much more expensive though)

In reality, at 60Hz, TN panel glass are essentially underclocked, often by an order of magnitude, as witnessed by those cheap 60Hz "1ms" TN monitors -- the exact same panels that have often been found in 144Hz monitors, just with much cheaper TCONs driving the panel glass.
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!

Q83Ia7ta
Posts: 761
Joined: 18 Dec 2013, 09:29

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by Q83Ia7ta » 30 Oct 2018, 19:24

a bit old twit https://twitter.com/australiamsi/status ... 2915722246
25 Oct https://au.msi.com/news/detail/176578e8 ... 21a008357f
29 Oct https://www.pcworld.idg.com.au/article/ ... australia/
"The MSI Oculux NXG251 gaming monitor will be available from December (via MSI's usual partners) at an expected retail price of $750." I guess it's Australian dollar (AUD). So expected price for AU will be ~ 532 USD.

darzo
Posts: 211
Joined: 12 Aug 2017, 12:26

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by darzo » 31 Oct 2018, 15:15

I wouldn't get too excited about 480hz monitors even by 2020. The 2080ti seems to be completely bottlenecked at 1920x1080 and CPUs are notorious for improving slowly. I have a 9700k overclocked to 5.2ghz with an AVX offset of 2 and in Call of Duty Black Ops 4 at the lowest settings I get spikes over 400 with lows near 240 fps, if not lower. Average fps I'm not sure reaches the 300s. So by 2020 either the next generation of graphics cards has a serious impact at a low resolution and low graphics settings or in the span of two years CPU performance improves by let's say 70%, which is highly unlikely. Maybe games start using the more cores and threads extremely effectively? Eh. Don't think there's much real basis to these refresh rate predictions, same as that "law" that claims over a certain period of time things become twice as good or whatever. What we're currently seeing is a slowdown. By 2020 both GPUs and CPUs should have undergone further shrinking and hence better improvements in performance if I understand correctly but 480 fps is very ambitious. And for actual 480hz monitors to exist there's not much waiting time in terms of development given the span of the timelines I've seen. We might have to hear something about them by the middle of 2019. I'd be surprised, pleasantly though.

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

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by Chief Blur Buster » 31 Oct 2018, 15:24

2020 is just for the first model, popularity will not be till much later in 2020s.

Remember how rare the first 120Hz monitor was in 2009, and how rare 240Hz still is today? Ditto. But the first model has to come somewhere.

And it will very close to 2020 (whether 2019, 2020, 2021 is simply nitpicking relatively speaking) -- but the 2020 prediction should be fairly accurate at least for the first consumer/prosumer 1080p 480Hz monitor.

Two huge reasons.

(1) Framerate amplification technologies for lagless & strobeless ULMB.

This is where new technologies such as Frame Rate Amplification Technologies has to come into play.

You know how Oculus Rift reprojects from 45fps to 90fps? That asynchronous timewarp algorithm? There is research in doing bigger ratios, such as 10:1 frame rate amplification ratios (e.g. 100fps converted to 1000fps) in a relatively lagless way and with less artifacts than today's reprojection / asynchronous timewarping.

A smaller amount of silicon is needed to amplify frame rates (via other methods) than to render completely from scratch, and so research is done in this. This is not your grandfather's Sony Motionflow with all the Soap Opera Effect artifacts, but a higher frame rate that begins to look like native frame rate.

I've written an earlier thread about this New term: "Frame Rate Amplification" (1000fps in cheap GPUs) and came up with a new term for this.

This is one of the steps we need for lagless & strobeless ULMB. It will probably trail ultra-high-Hz panels.

(2) At 480Hz and above, VSYNC ON and VSYNC OFF becomes almost the same lag even at 100fps

Keep in mind that VSYNC ON 60fps on a 480Hz panel has only a 2ms VSYNC penalty -- so 480Hz VRR panels can be practically as good as VSYNC OFF, as tested with the Blur Busters GSYNC vs VSYNC OFF lag tests. For 240Hz monitors, enabling VRR is almost but not quite as good as VSYNC OFF but CS:GO players still use VSYNC OFF instead of GSYNC, and scanout lag gradient is still 4.1ms.

With a 2ms scanout, even at lower framerates of a mere 100 frames per second -- the scanout lag gradent becomes virtually a non-issue, and can make VRR good enough (even at lower framerates) to use instead of VSYNC OFF. VSYNC OFF advantage is eliminating the lag issues of having to wait for the next scanout opportunity, or even interrupting the current scanout (tearline) -- and at 2ms refresh granularities nearly eliminate the lag-gradient (latency range / latency jittering / latency asymmetry) between VSYNC ON versus VSYNC OFF versus GSYNC methodologies, even for lower framerates such as 100fps.

Near zero scanout lag penalty. Near zero frame-transmission lag.
With faster frame-transmission of 1/480sec over a video cable, you eliminate the scanout lag (scanout lag = frame transmission lag). The VSYNC ON lag penalty is caused by scanout latency, and at 480Hz, the scanout penalty is only 2ms (1/480sec).

VSYNC ON and VSYNC OFF gets almost same lag
So when VSYNC ON and VSYNC OFF almost equalizes in input lag at any framerate (because of sheer fast scanout), the decision of which sync tech to use -- matters a hell lot less. Choose the way you prefer to play, rather than trying to fix lag.

Remember, VSYNC ON lag is the scanout lag / frame transmission lag / refresh cycle duration lag. This scanout lag goes down at higher Hz. Push it to near zero via sheer high Hz and ultrafast frame delivery -- and the lag difference of VSYNC ON versus VSYNC OFF becomes nil as we go towards 480Hz and 1000Hz

Microstutter becomes smaller at 100 frames per second at 480Hz
Also, microstuttering also will go down dramatically at 480Hz too, even at Windows desktop. 2ms granularities makes fixed-480Hz look almost like VRR for most low framerates (e.g. 50 to 100fps) since any beat-frequency / harmonic-frequency between framerate and refreshrate creates much smaller microstutter amplitudes -- e.g. 23fps@480Hz versus 24fps@480Hz versus 25fps@480Hz will look almost identical, unlike 23fps@120Hz versus 24fps@120Hz versus 25fps@120Hz. And the mouse-lock feel to display will be even better at 480Hz even for low framerates -- low framerates feel noticeably less laggy.
100fps@120Hz = very microstuttery (I clearly see it)
100fps@240Hz = less microstuttery (I can still see it)
100fps@480Hz = barely microstuttery (if still visible)

1000Hz starts looking equivalent to per-pixel VRR
It's like supporting VRR at multiple different framerates simultaneously on the same display. 1000Hz can essentially almost make VRR unnecessary, and make the display fully ambidextrous in framerate simultaneously due to the ultra-small refresh cycle granularities that essentially turns it into "a per-pixel VRR" behaviour rather than a "per-refresh-cycle VRR". Whether it's a browser scroll, or stuttery browser animation on fixed Hz. The microstutter amplitude of 23fps@480Hz is only 1/4th the vibrationwidth of 23fps@120Hz. All random low framerates start looking smooth simultaneously without need for VRR, so multiple animations running at different framerates (21fps in one browser window, 47fps in photoshop window, 63fps in a third browser window) all suddenly start looking much smoother simultaneously, because the refresh-cycle-granularity is so small, it ceases to become much of a microstutter issue. I've seen this phenomenon when testing at 480Hz that almost any random framerates no longer behave microstuttery at off-multiples. 24fps and 25fps videos (two windows side by side playing simultaneously) looked perfectly smooth. Yet 480 is not a multiple of 25. But it doesn't matter much because it's only a 2ms stutter in 40ms frame times (25fps = 40ms frames). So a 1/20th-amplitude microstutter is often below the noisefloor of visibility. This means you can play NTSC 24fps and PAL 25fps on the same fixed-Hz 480Hz screen at the same time, and both looks simultaneously perfectly smooth on my Zisworks 4K120Hz when it's in 480Hz mode (Sure, resolution is crap, but fast-motion looks completely non-stuttery in both video windows). This is a phenomenon that is beneficial. When a pixel is able to refresh anytime on demand to within a 1ms timing accuracy, the need for full-panel VRR diminishes hugely.

Tearing becomes less visible at 100 frames per second at 480Hz
Also, tearing is still visible on 240Hz monitors, less than 120Hz. Tearing at 480Hz will be even less visible.
100fps@120Hz = tearlines are visible for 1/120sec each
100fps@240Hz = tearlines are visible for 1/240sec each
100fps@480Hz = tearlines are visible for 1/480sec each
The briefer a tearline is visible for, the harder is to see them. I still easily notice tearlines at 240Hz in games for bright content -- dark vertical walls on bright backgrounds. Others can't. But, needless to say, tearing visibility is also an issue.

All this combined, this helps you even at only 100 frames per second. Once 480Hz is cheap enough to add. Being the world's first website to test 480Hz (and it still sits on my desk next to other monitors) -- I have discovered many unexpected benefits -- like playing PAL 25fps and NTSC 24fps simultaneously side-by-side in a stutter free way -- that's tantamount to behaving like per-pixel VRR simply by sheer near-zero refresh cycle granularity, eliminating refresh cycle granularity from becoming a stutter factor. 1000Hz mice were dismissed as hooey at first until most grudingly agreed it worked. 480Hz is also cheaper to pull off than a FALD backlight. 480Hz displays are not hooey, for many reasons I've explained above. It benefits low frame rates too.

Imagine replacing GSYNC, FreeSync, VSYNC OFF, VSYNC ON, Fast Sync, Enhanced Sync, LightBoost, ULMB, with just simply sheer Hz
Once you've got ultra-high Hz, a hell of a lot of display motion problems are solved simultaneously in a virtually lagless way. Who cares about the sync tech or blur reduction tech once you go past a specific refresh rate. (For the blur reduction part will still require high frame rates -- or will still need strobing tech to eliminate motion blur in low-framerate material -- but for everything else, VRR is easily replaced via ultra-high-Hz). Heck, imagine VSYNC ON and VSYNC OFF becoming identical in input lag when a display's frame granularity approaches ever closer to 0ms. Even if it won't happen quickly, but it is a worthwhile journey in the Amazing Journey To Future 1000Hz Displays. I will probably be writing a sequel to that article, because I have never (yet) written how 1000Hz behaves like per-pixel VRR. And I haven't clearly written about how the difference between VSYNC ON and VSYNC OFF and VRR converges and disappears as refresh granularities approach 0. But it does and I've visually confirmed that behaviour, as has many researchers of 1000Hz+ displays. There are many additional benefits of 480Hz+ that now belongs in a new followup article. Consider this forum post a taster.

It'll still be a niche sure, but so is 240Hz. (And even 1000Hz mice was a niche for a long time till they became a widespread standard feature in gaming mice). For years, this was pie-in-the-sky stuff. But an indie single-handedly added 480Hz to an off-the-shelf 4K60 LCD panel -- and this is no longer unobtainium. Now the engineering world just has have to fill-in the missing technology pieces (e.g. for GPUs to gain more frame rate amplification tech) to allow manufacturers to properly commercialize the benefits of this.

Do not underestimate benefits of 480Hz operation, even for low frame rates.

Further reading for other readers not aware of our ultra-high-Hz research:
- Our 480Hz Tests
- Blur Busters Law And The Amazing Journey To Future 1000Hz Displays
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!

darzo
Posts: 211
Joined: 12 Aug 2017, 12:26

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by darzo » 31 Oct 2018, 22:09

Have you spoken to anyone about producers being willing and able to provide these monitors within the next two and a half years? Not a very tactful question but I wonder whether you know more. I also keep seeing you mention framerate amplification as something realistic (which makes me wonder why we would need to spend a lot of money on graphics cards at that point, i.e. the current framerate amplifiers). Is that in any stage of becoming a reality or just theoretical at this point?

mello
Posts: 251
Joined: 31 Jan 2014, 04:24

Re: Touch Taiwan 2018 - Any more news on 240hz 0.5ms?

Post by mello » 01 Nov 2018, 10:11

Great post Chief, thanks again for more information about the benefits of high hz displays ! :mrgreen:
darzo wrote:Have you spoken to anyone about producers being willing and able to provide these monitors within the next two and a half years? Not a very tactful question but I wonder whether you know more.
I believe that this prediction is based on progress and advancements that were made in last years (and that zisworks already made it before any major manufacturer), so ~2020 and 480hz @ 1080 seems like a real possibility.
darzo wrote: I also keep seeing you mention framerate amplification as something realistic (which makes me wonder why we would need to spend a lot of money on graphics cards at that point, i.e. the current framerate amplifiers). Is that in any stage of becoming a reality or just theoretical at this point?
It has already been done in Oculus. But i have also asked similar question before, when Chief first wrote about FRAT. For me it looked like it might disrupt a GPU market in major way, because you might be able to get 100fps>1000fps on mid range GPUs when the tech finally arrives. So i had similar question, why anyone would want buy high end GPUs ?! From what i understand GPU power and progress in that area is still highly desirable, regardless of FRAT. But Chief would need to go more into detail about that.

Post Reply