First Impressions of the Dell UP3218K "8K" Monitor

[Baasha]

Just got my Dell 8K monitor on Friday evening and set it up.

I've had this panel for just over a day so take my comments with a grain of salt.

The 8K image is stupendous - it is so crisp and clear that I have never seen anything so sharp and beautiful. Gaming on it is absolutely fantastic - the main rig crushes most games and goes well beyond 60fps so it's buttery smooth and great to look at. It is a glossy screen and bezel is nice and thin. The buttons are on the bottom - the fit and finish is superb.

This is a great monitor for content-creation - which I will be doing quite a bit of, as well as gaming. For gaming purposes, if you're considering this monitor and plan to use it at 8K 60Hz (doh!), two Titan Xp will NOT cut it - getting even 30fps in newer games will be challenging. 8K is 33.2MP which is 16x 1080P or 4x 4K.

Some things I noticed: there is some backlight-bleed on the lower left of the monitor - irritating when there is a complete black background but unnoticeable otherwise. This panel is pixel-perfect - ran through the LCD screen tests and not a single dead/stuck pixel so I'm stoked.

Some snips:

@ 100% DPI scaling:

@ 300% DPI scaling (recommended):

dat monitor real estate tho:

 

[Mr Evil]

Very nice indeed!

How close do you have to get to be able to see individual pixels? I found that with a 32" 4k screen, I have to move back to about 1.5m before I can no longer locate a single black pixel on a white background. That made me think it would be beneficial to have an 8k display, although I would rather wait until GPUs and interconnects have advanced enough to do it at 120Hz (and without breaking the bank).

 

[mohit9206]

And to think i am still on 1600x900..

 

[guachi]

I do a lot of photo editing (bad photo editing. but, still, I edit photos) and my camera doesn't even have a 33.2 MP sensor.

Though at least you could show the photo 1:1 and still have space for control windows, I suppose. I will say that 4k for editing is such a boost over 1920x1200 I'm glad I upgraded.

 

[amenx]

I would love to see a test involving a 100 people sitting 2' away from 32" 4k and 8k screens and asked to spot the difference. Maybe a few can, but I'll bet the vast majority wouldnt.

 

[tential]

I would love to see a test involving a 100 people sitting 2' away from 32" 4k and 8k screens and asked to spot the difference. Maybe a few can, but I'll bet the vast majority wouldnt.

After going from 1080p to 4k, you can.
The resolution and screen real estate clarity is night and day. Would 100 users who don't need an 8k screen notice and just general PC users? Probably not.

For anyone who can buy this monitor who is a high end PC user? Definitely. It's just SO MUCH screen real estate and it does a lot. Resolution is so important, I'm excited for 8k actually.

It's one of the first things I thought after spending 5 days with my monitor is "I wish we had 8k...."

 

[amenx]

After going from 1080p to 4k, you can.
The resolution and screen real estate clarity is night and day...

Not sure I understand your point. Tbh, talk of res without referencing screen size (and viewing distance) is pretty much useless. Of course going from 1080p to 4k would be night and day (if 27-32", but not 'night and day' on 5" screen phone) . 4k to 8k on 32"? The higher up you go in res, diminishing returns sets in for a given size. Again, only way to beat that is increasing screen size. I am sure I would be very impressed at diff between 4k vs 8k on a 100" screen at 8' away, but quite doubt that on 32" @ 2'.

 

[Valantar]

Though at least you could show the photo 1:1 and still have space for control windows, I suppose.

What's the point of a 1:1 display if the pixels are so small you can't make out any detail? I suppose moving your head closer is quicker than zooming, but more straining on the eyes too. Still, 8k for photo editing doesn't sound like it will have any real use outside of 40"+ displays.

After going from 1080p to 4k, you can.
The resolution and screen real estate clarity is night and day.

Not a comparable scenario. Returns diminish significantly as DPI rises. There's a reason 4k smartphones haven't become a thing yet. 1080p over ~24" at 2' gives you quite easily distinguishable pixels. 4k? No. So you start out with radically different playing fields. When most people can't make out individual pixels on a 4k display (which the same people could easily do on a 1080p one), of course they would struggle to tell the difference at 8k.

I found that with a 32" 4k screen, I have to move back to about 1.5m before I can no longer locate a single black pixel on a white background.

That sounds like an awful benchmark for what you're trying to test - our eyes are very well attuned to picking out single details in plain fields of color, which isn't really a realistic use for a monitor. Try displaying a single-pixel black and white checkerboard pattern, and see when it stops looking like a pattern and starts looking like a grey/mottled grey field, or do the same with alternating black and white 1-pixel tall/wide lines horizontally/ vertically. Or a single diagonal line and see how close you have to go for aliasing to show. That would be more relevant of a benchmark.

 

[Mr Evil]

...That sounds like an awful benchmark for what you're trying to test - our eyes are very well attuned to picking out single details in plain fields of color, which isn't really a realistic use for a monitor. Try displaying a single-pixel black and white checkerboard pattern, and see when it stops looking like a pattern and starts looking like a grey/mottled grey field, or do the same with alternating black and white 1-pixel tall/wide lines horizontally/ vertically. Or a single diagonal line and see how close you have to go for aliasing to show. That would be more relevant of a benchmark.

What I'm trying to test, is the point where increasing the resolution no longer makes a visible difference. For that, the most sensitive test is appropriate. But ok, for a more realistic scenario, I tried comparing text (whatever the default size in Notepad is, at 125% DPI scaling) with antialiasing enabled and disabled, and I can tell the difference at pretty much any distance where I can still read it, which is further than I can see single pixels.

 

[tential]

If most people can't distinguish pixels at 4k, that's too many people. 8k minimum, 16k for the elite.

 

[Valantar]

What I'm trying to test, is the point where increasing the resolution no longer makes a visible difference. For that, the most sensitive test is appropriate.

No, it isn't. Why? Because with the most sensitive test, you're testing an extreme edge case, which is then by default not applicable to the vast majority of ordinary use cases. You're defining "visible difference" as "any visible difference in any situation" which is not a practical definition for any normal monitor use. Not to mention that a whole host of variables besides resolution affect this: brightness, contrast and ambient brightness to name a few. A brighter monitor might burn out the black pixel in your eyes, or too dark ambient lighting might do the same. Or are you testing across a range of display and ambient brightness settings? Also: confirmation bias and knowing where the pixel is/is supposed to be is bound to make you see it more clearly than if it were placed randomly. Following a known object with your eyes and spotting an unknown one are not the same.

 

[ioni]

talk of res without referencing screen size (and viewing distance) is pretty much useless.

For smartphone screen sizes, sure. There's really no need to go beyond 720p for a 5" screen. But for PC monitor sizes, res is res. You don't need to put it into any extra context.

Not a comparable scenario. Returns diminish significantly as DPI rises.

DPI has little to do with it. The extra screen real estate you get from the res bump is night and day enough. Whether you scale the screen size up so it has the same DPI or keep the screen smaller to get a higher DPI.

 

[Valantar]

DPI has little to do with it. The extra screen real estate you get from the res bump is night and day enough. Whether you scale the screen size up so it has the same DPI or keep the screen smaller to get a higher DPI.

Screen real estate only increases if you don't use any form of scaling. I'd love to see you use a 24-30" 8k monitor for any practical purpose at 100% scaling from normal viewing distances (~2'). Again: at a certain point, diminishing returns start kicking in.

 

[guachi]

For smartphone screen sizes, sure. There's really no need to go beyond 720p for a 5" screen. But for PC monitor sizes, res is res. You don't need to put it into any extra context.

Text crispness is incredibly different on a 5" screen between a 720, 1080, and a 1440 display.

Oh the other hand, I can't tell the difference when playing a video.

 

[Mr Evil]

No, it isn't. Why? Because with the most sensitive test, you're testing an extreme edge case, which is then by default not applicable to the vast majority of ordinary use cases...

You seem to be trying to tell me that I should be satisfied with good enough. Nonetheless, as long as a monitor can be distinguished from one of infinite resolution, then there is room for improvement, and advances in technology will make it practical eventually.

...a whole host of variables besides resolution affect this: brightness, contrast and ambient brightness to name a few...

I expect most people will have their monitors set up in their preferred conditions at home (that's harder to control at work of course). I'm not expecting 10 sig figs of precision here, I just want to know what experience people have.

...confirmation bias...

I'm well aware that I am a fallible human, but I wasn't aware that I was supposed to do ABX testing for simple forum posts now.

Apologies to Baasha for derailing the thread with my question.

 

[Baasha]

The monitor absolutely brilliant. The hard part is that it really comes down to this - I have a 5K monitor (HP Z27Q) that's 27" and 217PPI.

Does this 32" 8K make the images so much better that it warrants its $5K cost? That's iffy...

The thing that irks me the most is this monitor requires TWO (2x) DP 1.4 cables and doesn't have DSC - if it did, it could do 8K @ 60hz with one DP 1.4 cable and all the issues with MST would disappear.

The image is second to none - the quality, crispness, and high-resolution are so incredible. I really hope they make an 8K OLED panel relatively soon that requires just one DP 1.4 cable for 8K @ 60Hz.

 

[nvgpu]

It's a pro monitor, any lossy compression is unacceptable when you need colour accuracy. And there will not be single cable DP 1.4 driving 8K 60Hz, you need a newer DP standard, 1.5(?)/2.0(?) with enough bandwidth to drive with a single cable like HDMI 2.1 48G cable.

 

[Baasha]

It's a pro monitor, any lossy compression is unacceptable when you need colour accuracy. And there will not be single cable DP 1.4 driving 8K 60Hz, you need a newer DP standard, 1.5(?)/2.0(?) with enough bandwidth to drive with a single cable like HDMI 2.1 48G cable.

No, 8K @ 60hz is doable with DSC. In fact, the upcoming Asus 4K 144Hz monitor has DP 1.4 with DSC iirc. Dell should've really gone with that - this monitor is essentially 2 panels acting as one (like Surround) and hence it requires two cables to run 8K @ 60hz.

 

[Mr Evil]

It's a pro monitor, any lossy compression is unacceptable...

No, 8K @ 60hz is doable with DSC...

Display Stream Compression is described as "visually lossless", which is marketing-speak for lossy.

 

[Excessi0n]

What I'm trying to test, is the point where increasing the resolution no longer makes a visible difference. For that, the most sensitive test is appropriate. But ok, for a more realistic scenario, I tried comparing text (whatever the default size in Notepad is, at 125% DPI scaling) with antialiasing enabled and disabled, and I can tell the difference at pretty much any distance where I can still read it, which is further than I can see single pixels.

If you want the most sensitive possible test, play around with this aliasing visibility test. The human eye can pick up aliasing/jaggies from far beyond the point where it's impossible to distinguish between individual pixels. With my 23" 1080p monitor I can see the jaggies from the end of my hallway, 25 feet from the monitor. With any of the test settings.

That's the worst case scenario, sure, but it shows that we have a looooooooooooooooong way to go before looking at a monitor is like looking through a window.

 

[imported_bman]

Display Stream Compression is described as "visually lossless", which is marketing-speak for lossy.

It looks to be pretty minimal from watching the video above where they use a 16x diff to show the effects of compression.

The above video at the 15:00 explains how 'visually lossless' is defined.

 

[guachi]

If you want the most sensitive possible test, play around with this aliasing visibility test. The human eye can pick up aliasing/jaggies from far beyond the point where it's impossible to distinguish between individual pixels. With my 23" 1080p monitor I can see the jaggies from the end of my hallway, 25 feet from the monitor. With any of the test settings.

I have 20/400 vision and, while lying in bed, I don't have my contact lenses in when I use my phone (2560x1440) before I go to sleep.

At a distance of about a foot all the text on my screen is a total blur. And yet at arm's length (about two feet) I can still see the aliasing on the test pattern even if I can't tell there is any text on the screen.

 

[fleshconsumed]

As gorgeous as that screen is, I'd rather have 5K 27" OLED at a reasonable price. I have 4K 27" IPS and it's borderline "retina", I think 5K should do the trick, OLED would be icing on the cake. A screen like that would be my "forever" screen.

 

[Valantar]

You seem to be trying to tell me that I should be satisfied with good enough. Nonetheless, as long as a monitor can be distinguished from one of infinite resolution, then there is room for improvement, and advances in technology will make it practical eventually.


I expect most people will have their monitors set up in their preferred conditions at home (that's harder to control at work of course). I'm not expecting 10 sig figs of precision here, I just want to know what experience people have.


I'm well aware that I am a fallible human, but I wasn't aware that I was supposed to do ABX testing for simple forum posts now.

Apologies to Baasha for derailing the thread with my question.

I'm not saying you should be that thorough, I'm just trying to warn you that an approach like that can easily lead you to audiophile-level money wasting chasing ever less tangible improvements A reality check once in a while never hurt anyone, after all. But I don't doubt for a second that there is a noticeable difference even between, say 27" 8k and 27" 16k at 2' - our ability to make out lines, parallels, unevenness and so on is quite impressive. It's just that once we start getting to the level where weirding out our eyes'/brains' image processing is the only way to tell the difference, I'd say we're way past useful improvements.

(As an interesting side note, in the (great!) Silo novel series by Hugh Howey, the resolution of the displays in the reality-altering helmets put on people sent out of the silo to "clean" (read: die by exposure to an extremely hazardous environment) is something like 30000*8000 at a few inches, and is described as indistinguishable from reality. But then again, it's a work of fiction, plus those people have been raised for generations in an underground silo/bunker controlled by absolutely insane people, and might not be used to high resolution gaming monitors. Still worth a read, though, even if I just partially spoiled it for you )

 

[Fir]

Looks nice, will wait for price to drop.
I remember when the 4K Dell 32" was $2500.
Now it can be found for under half that.