Potentially low MPRT on many 4K and 8K Samsung QLED TVs (lots of RTINGS BFI Test image comparisons)

[JimProfit]

Hello everyone, I'm new to the forum but active on many display-tech-focused platforms.

While browsing the table tool on RTINGS, hoping to find out which TV on the market has the lowest MPRT via BFI/strobing, I noticed that two recent Samsung LCD models clearly stood out in the "BFI Picture" category.

The higher-end QN95D and the upper-mid-range QN90D from 2024, both display a very clear test picture (very thin fringe of blur on the sides of the logo elements, and just a tiny bit of ghosting, being LCD).
Both sets tick a lot of boxes as far as the "best overall gaming tv™", being very bright in both HDR and SDR, having great contrast by LCD standards due to their numerous dimming zones, and I assume having access to the low-lag frame-gen that Samsung also uses on their QD-OLEDs.
I tend to favor the QN90D as it supports strobing at both 60hz and 120hz, while the QN95D only does 60hz. This feeling is further reinforced by the fact that the picture appears much brighter in the BFI test results.

However the possibility that these sets might have close-enough-to-CRT-motion on a large panel is tantalizing indeed.
I know ultra-short backlight pulse LCD can match or exceed CRT motion clarity: my 3D Vision monitors (PG27V and PG278QR) both do that via ULMB, and I was also made aware by Ariel, the guy from the Youtube channel "Plasma TV for Gaming" that some Sony TVs from the early 2010s have very low MPRT via "Motionflow Impulse" strobing.

I combed through the RTINGS table tool and inspected every "BFI Picture" screenshot to see which is the absolute least blurry/artifact-free.
It is absolutely clear that those Samsung sets wreck most current options.
I own both an LG G1 and GX, the GX being the gold standard of OLED motion clarity (MPRT roughly equivalent to plasma level) and their BFI Picture test are clearly inferior to the Samsung.
While I have basically no chance to find one of those Motionflow Impulse sets from way back that have CRT/ULMB level MPRT, I have however found the review of a Sony model from 2015 on RTINGS, the X850C on RTINGS, that does have a very impressive test result.
In subsequent models Sony's BFI performance took a serious nosedive. Which leaves me to believe that the X850C might have been the last/best Sony TV that had the Motionflow Impulse setting, albeit under a different name.
If that hypothesis is true, and the RTINGS pictures are accurate, it would mean that the QN95D and especially the QN90D in 120hz, would beat these absolute champions of motion clarity in TVs.

OLED will never go beyond Plasma-level motion (edit: at 60hz), and that's only if LG get's their heads out of their asses and bring back their rolling-scan tech from the 2020 and 2021 OLEDs series.
With backlight-strobing LCDs and lots of dimming zones for contrast, we could in theory emulate the CRT qualities in modern living-rooms, all the while with above-SDR brightness. What a compelling possibility.

Anyway, I have no means to go test these sets in a store or anything, I might pick a QN90D if I find one on sale, however this post is mostly to raise awareness (not to be misinterpreted as shilling).
What do you guys think? Have any of you guys seen the motion performance on these sets or their previous years counterparts (which tend to be no slouch either, some of them).

[Chief Blur Buster]

OLED will never go beyond Plasma-level motion

Um.... "Depends on OLED Hz"

My desktop 480Hz successfully has less motion blur than my old Pioneer Kuro plasma display, since 2ms MPRT is less blur than the 3ms phosphor, but only for 480fps content. But I can also now run a new CRT simulator to reduce motion blur by up to 87.5% for 60fps content running at 480Hz.

We're migrating to BYOA. Bring Your Own Algorithm today. With my open source shaders.

We're implementing it into Retrotink 4K now as we speak, I'm working with Mike.

Motion Blur Reduction Video Processors

Future video processors will let you beat certain lineitems of TV firmware BFI, simply by piggybacking on brute Hz. In fact, Retrotink box-in-the-middle 4K BFI is brighter and lower lag than the LG built-in BFI too! (We found ways to do that). The nice thing is that with Blur Buster's migration to the new approach for 2030s, is that we only need generic Hz. In the future, I can reduce almost 90% of 60fps motion blur with a box between your PC and display.

Vendors reading this, I've worked with video processors and line doublers since year 2001. Key Digital LEEZA, TAW ROCK, Runco, and a Faroudja Fli2200-based PCI card, plus I created an open source 3:2 deinterlace algorithm for dTV/dScaler in year 2001. I can now motion blur reduce via box-in-middle approaches. Vendors can reach out to me at blurbusters.com/about/contacct

Anyway, I have no means to go test these sets in a store or anything, I might pick a QN90D if I find one on sale

If you purchase LCD strobing, make sure you check the strobe crosstalk (top/bottom may have more double image artifacts than center). You often can solve that using refresh rate headroom (like Large Vertical Total tricks, or NVIDIA-quality strobe tuning, etc).

[JimProfit]

Um.... "Depends on OLED Hz"

Indeed, I was referring to the likelihood of one day having more than the GX/CX 25% rolling scan at native 60hz resulting in a ~4ms MPRT.
At 120hz, my GX/G1 would reach 315hz equivalent with Motion Pro High.
And if resolution matters, then a native 480hz OLED is a moot point since they're currently stuck at 1440p max.

Of course OLEDs definitely have their place, for movies and slower moving content you can't beat that contrast, however I'm holding out on upgrading until the panels have a dramatically higher refresh rate than the signal they can take, for the sole purpose of built-in, low-lag frame-gen.
The Samsung Game Motion Plus mode looks promising, but it so far it doesn't go beyond the native refresh of 144hz (same as classical laggy motion interpolation).

Nonetheless, for SDR content/retro stuff stuck at 60hz max, I could see myself living with the downsides of LCD, provided I get a close approximation of a giant CRT in terms of motion.
My ULMB monitors look stunning in motion, and are at the very least much larger than my FW900, the problem lies with the terrible contrast ratio (edge LED without dimming zones made even worse by the craptacular anti-glare coating, the picture quality itself is terrible considering my monitor sits right next to my G1).
Also, while I've been able to hack it to run ULMB at 60hz thanks to CRU, there's no feasible way to use anything other than a PC at that refresh, and the 1440p resolution isn't console friendly anyway.
But I would be so happy with a larger, 4K HDR version of that, with lots of dimming zones to improve contrast ratio to acceptable levels.

With that in mind, do you reckon those "BFI Picture" test results on RTINGS have any merit? Can their testing protocol truly capture the motion clarity with consistency between TVs?

Perfection is not for this world and I've resigned myself that I may need to line every wall in my house with screens if I want the best possible display for every circumstances

And I'm of course looking forward to the future, I'm delighted to hear you're working with Mike Chi.

[Chief Blur Buster]

With that in mind, do you reckon those "BFI Picture" test results on RTINGS have any merit? Can their testing protocol truly capture the motion clarity with consistency between TVs?

Oh yes, for screen center, the pursuit camera images are accurate. A problem however, is the top/bottom edge may not have the same motion clarity. Not all of the screen is measured, but they measure motion blur more accurately than the majority of reviewer sites.

Perfection is not for this world and I've resigned myself that I may need to line every wall in my house with screens if I want the best possible display for every circumstances

With a fast 1000Hz+ OLED, one can do display emulators, and reduce the count of screens. You can temporally simulate slow LCD GtG on an OLED, or plasma on OLED, or CRT on OLED, etc. I released a CRT simulation shader recently at www.blurbusters.com/crt as a proof of concept.

[JimProfit]

Thank you so much for your input.
I'm comforted in the idea that those are worth investigating in the here and now. I don't know when OLEDs will come with native 1000+hz panels (and again, I assume the input signal will be lower for a good long while, especially at higher res), but obviously, as long as frame-gen keeps improving, it is the absolute way to go in order to preserve HDR peak brightness.

So, I tried your strobe crosstalk test on my PG27VQ, I do get some artifacts in the lower part of the screen, but it's a small price to pay for the CRT-level clarity afforded to me by ULMB.
So if I understand correctly, as long as the crosstalk artifacting on top and bottom does not completely ruin the motion, I would be correct in my belief that those 2024 Samsung TVs are the lowest MPRT HDTVs on the market in the here and now, based on their BFI Image test results for RTINGS then? The closest thing to a giant 4K CRT we can actually buy today?

For reference, I'm adding the picture for the 2024 models QN90D, QN95D, and the 2020 LG GX, the latter being the lowest MPRT 4K OLED at 60hz and 120hz w/ BFI set to high (x4 motion clarity multiplier).
I'm not sure if the GX and QN90D results are captured at 60hz or 120hz. QN95 appears the clearest, but is a bit dim and BFI is limited to 60hz, so I'm leaning towards the QN90 and hoping the test result is representative of 60hz BFI...

[Chief Blur Buster]

They are certainly better than average looking;

I can't say for certain if they're definitely the best or not. But it's still better than average.

I've seen better on the XG2431, but large-panels are often slightly slower than small-panels.

[JimProfit]

Yeah, I really combed through all the BFI picture tests in the table tool. Not only are those two the best results for TVs currently on the market, but also of all the past TVs with older test protocols. I simply could not find one that beats those.

I mostly concentrated my analysis on the thickness of the fringe of blur on the sides of the letters. And on LCD models, I paid attention that the ghosting and double-image artifacts were tolerable.
In which case the 60hz-only QN95D is the best, but I wonder why it shows such a dark outline all around the logo elements, that also appeared on the last-gen QN95C's BFI test, which leads me to believe it is an artifact of those TVs motion processing.

I'm still leaning towards the QN90D, hoping that the test image was done at 60hz and that the 120hz mode closes the gap with the QN95D while maintaining a higher brightness. I can definitely wait to see the results on next year's models of course.

As I said I of course look forward to 1000hz 4K OLEDs, but it might be a while before they're available to purchase. And because 1000fps is pretty much never happening in AAA games, we'll need to rely on frame-gen for these extreme refreshes to be of any use.
Which, I hope TV manufacturers will have solid frame-gen built-in, because not every system supports it or at least not to the extent of reaching 1000+ fps.
Speaking of which, I love lossless scaling, and looking forward to having a newer GPU that does DLSS frame gen well (I'm on a 2080TI waiting to upgrade), however I've seen some pretty disastrous results on some occasion, so it will also be a few years before frame gen becomes the universal solution for high-refresh panels. The algorithms need to improve in conjunction with the increasing refresh rates.

By purchasing a very bright LCD 4K TV with very low MPRT in the present, I solve a few problems for myself:
-I don't downgrade screen size or resolution by targeting a 1080p 480hz monitor that'll have to rely on frame gen (which can on occasion look terrible).
-I can increase the motion clarity of every past system that doesn't/never will have frame gen, i.e an especially perfect solution for everything from the past that is SDR and 60hz max.
-I have enough HDR fullscreen brightness that I can get better-than-SDR highlights even when I leave backlight strobing set to max (which on my GX and G1 is basically sacrificing all available brightness to reach barely-above-SDR image and plasma-level motion clarity.
Or, for older games, I would probably have enough brightness for both low-MPRT strobing AND CRT filters, that I normally never use because they dim the picture and BFI takes the priority.

Edit: regarding the strobe crosstalk breaking down motion clarity on the top and bottom of the screen, can it be a function of screen size? Like the problem would be worse on a 75" TV than on a 55" of the same model?

[JimProfit]

After looking at the BFI graph, I believe with some certainty that the GX and QN90D BFI picture test were conducted at 60hz and should therefore both be clearer at 120hz.

Also, I had dismissed the 8K models from my analysis, but the QN900A/B/C/D all display exemplary BFI test results.
Only the QN900A and QN900B strobe at 120hz though.
The former retains a lot more brightness, the latter has far superior contrast.
It's really hard to decide and a bit infuriating that newer high-end models have very dim 60hz-only strobing.
I'm again in that GX/G1 situation where a crucial features seems to go the way of the dodo and I resort to buying discontinued models instead...

While I don't care for native 8K in the absence of content, it is relevant to retro gamers, as it is the first resolution to be an integer of ALL major standards: 240p, 480p, 720p, 1080p, 1440p and 2160p...
One could imagine using a scaler to bring a 480i/p signal to 1440p and let the TV bring that to 8K...With low-MPRT and CRT mask filters, when I bring up the giant CRT emulator, now we're really talking....

Edit: Pictures for reference (assumed to be at 60hz)
Edit 2: From what I understand, the QN900B has more settings for local dimming, and it's fairly safe to assume that the BFI test was conducted with it set to Local Dimming High, perhaps that is why the BFI graphs are affected on some models (tiny hills instead of sharp peaks). I emit the assumption (because assuming is the best I can do at the moment) that, with a more modest local dimming setting, we could have more brightness when strobing is on, perhaps matching or beating that Q9N00A which, with its sole local dimming setting set to On, still has much lower contrast in comparison to the QN900B with Local Dimming on High... What a rabbit hole...

[Nocta]

[...] I released a CRT simulation shader recently at www.blurbusters.com/crt as a proof of concept.

Just wanted to say thank you and congratulations. This is finally a reality and it's incredible. We just entered the post-CRT era thanks to your hard work!

[Chief Blur Buster]

[...] I released a CRT simulation shader recently at www.blurbusters.com/crt as a proof of concept.

Just wanted to say thank you and congratulations. This is finally a reality and it's incredible. We just entered the post-CRT era thanks to your hard work!

You are welcome!

I'm going to release more display simulator shaders including plasma TV simulator and LCD simulator shaders, over the years -- as part of the Blur Busters Open Source Display Initiative. This is to help prod the ecosystem of improved refresh cycle processing without waiting for display manufacturers.