Are our screens REALLY 1080p? Check out this picture.

Since getting my XPS 17 with the 1080p screen, I've noticed I can easily see pixelation, especially in text. Even glancing at the Windows Start button, the edges seem quite a bit more aliased than they were on my old (larger) 1080p HP.

I initially just chalked this up to contrast tweaks or imagination. But I never suspected the screen was faking its resolution... Until now.

First, do this test. http://www.lagom.nl/lcd-test/clock_phase.php

If the huge square is pixelated and differently colored (or shade of gray) than the reference images, that's a big problem. For my screen, the square appears with a light beige/yellow color and is clearly pixelated at native resolution and zoom (you don't even have to squint to see that). I took a screenshot and blew it up in an editor, and sure enough, the color that the monitor displayed (light off-white/beige/yellow) doesn't exist anywhere on the page. This is evidence of subpixel interpolation (a type of trick some shady monitors use to 'fake' their native resolution).

This suspicion first reared its head when I was doing some dithering. Check out this image, which has a nice natural gradient in it. I dithered it using ordered patterns to soften the borders between the gradients. I was shocked to see a bright, gaudy rainbow suddenly pop out of the gradients. These large, discolored rainbow rings appeared at each dithered border. Do you see the rainbow? Bright yellow/orange/green bands where the colors should all be relatively similar. Now zoom in with your browser (ctrl +) or save the picture and zoom in with any program. Voila! You'll see the weird rings vanish, leaving only the original intended gradations. No ridiculous alternating bright colors.



If you can clearly see amplified rainbow-colored ring borders at the gradations (at native resolution), your monitor is likely faking its resolution and doing some hardware scaling trickery. Best of all, notice how the bands and colors change dramatically when you slowly drag the image (at original size) horizontally across the screen. Try it right now: take this browser window and slowly drag it to the left or something. Notice how these jarring discolored "rainbow fences" drastically diminish when you change your screen resolution (like to 1600x900), or when you resize/zoom the image. Or when you look it it with a real 1080p screen.

Now, I've seen this maddening dithering/resolution phenomenon before in the form of the PenTile screen used to "cheat" the resolutions on some older smarphones.

Call me crazy, but it almost seems as if Dell is trying to pull a similar stunt with these laptops. If it does use something like a PenTile matrix, that means the effective resolution could be something closer to 1574x886 (yes, potentially less than that 1600x900 monitor in your basement).

Needless to say, this makes doing any pixel art/dithering work impossible on this laptop at the native resolution we paid for. Am I overlooking something? Is there an easier answer here?

Colors shouldn't appear out of thin air when they're not there. Text and sharp graphics shouldn't appear aliased. Dithering shouldn't produce sharp rainbows of color. Something's terribly fishy here, and I smell a resolution issue...

 

With my 1080p AG screen i find the AG coat on the screen responsible for the ' pixelation', but they are 1080p resolution.

 

I'm not talking about the subtle matte 'grain'. I'm talking about heavy aliasing and color artifacts indicative of subpixel interpolation (i.e. screen not really RGB 1080p).

Can anyone with an XPS with 1080p screen take a look at the first picture and try this? This is an absolutely remarkable finding that needs some attention.

 

I see the rings in the picture and I am no expert but the screens are still 1080p, the quality of the screen is debatable though.

 

Review Clock and phase - Lagom LCD test.

 

hmmm I get a similar effect in the smoke on the startup screen of Skyrim. Yet when I view it on my 32" 1080p LCD TV, there is no banding and the smoke is perfect

here is a zoomed-in screen shot from Skyrim on my L702x



and here is a photo of my LCD TV (sorry about quality but trust me, the smoke is sharp as a thistle)



Both screens are set to 1080p. Maybe there is a setting I can enable to make the smoke look better on my L702x? I might be missing something obvious

edit: uh oh I think my pics are broken

 

Looks fine on my L702x, definitely 1080p

 

Thanks for the feedback.
VERY helpful link with the monitor test. This certainly proves my point -- it's clearly light yellow/beige and (just discernably) pixelated, NOT the flat gray it should be. This is at native resolution. This is the proof, right here.

It's very interesting that the monitor shows a beige/yellow discolored image partway between the left and the right "what it should and shouldn't look like" pictures. The picture should be gray and flat like all the other pictures in the demo. I took a snapshot of the screen, zoomed in, and analyzed it -- sure enough, once I zoomed in, the light beige color vanished, and it turns out every pixel is either white or dull gray, and there is NO yellow/beige/cream color in the lattice whatsoever. The monitor is clearly at fault here -- I can even make out the tiny checkered lattice, countless tiny but visible dots, though not bad enough to look like an uneven quilt like it does on a genuinely 'non-native' resolution. The only test the monitor fails is the resolution test, sure enough. But that test seems to be designed only to tell you if you're running at native resolution -- not if that resolution itself is accurate. Like I was saying, if the picture I linked to in the OP shows the banding, and/or the test shows a bright beige sandy screen instead of the dull, perfectly flat gray it's supposed to, the monitor is cheating with its native resolution. It isn't 1920x1080, even if it registers as such in Windows.

The important part is, these vivid color bands (or the bright beige) DO NOT EXIST in the actual images even though they show up on the monitor (in native resolution and 1:1 picture zoom). Sure enough, when you zoom in or out even the tiniest bit, or change the monitor's resolution, they disappear. The colors revert to normal, even though the zoom is way off.

Try the test linked above, then try this with the picture I posted in the first post: Move the picture very slowly, horizontally across the screen, left to right (or right to left). You should see the bands change colors and positions dramatically as the picture is moved. When you place it down somewhere else, it looks like a completely different picture! These are all clear signs the screen is not 1080p. The OS and all graphics render at 1080p and the monitor happily accepts such input, but it is physically incapable of displaying all of the pixels, and it mashes some together at the subpixel level (RGBG?).

Don't get me wrong, all the pixels are "rendered" by the operating system, yes. Otherwise you wouldn't be able to select and run Windows or any software at this supposedly 'native' resolution. The point is, even though they're rendered at a software level, they're not actually physically displayed -- many close neighbors to each pixel are being mashed together behind the scenes in color and resolution, and this is what you get. It's especially obvious when looking for high contrast edge aliasing or when working with pixel-perfect dithering. In other words, any kind of pixel-level design, dithering, or shading work.

Once you appreciate the significance of the phenomenon observed in the first picture and in the test linked by Darkstone, you'll realize what I mean.

PS: Many people purportedly can't tell the difference between DVD and Blu-ray resolution (480p and 1080p) movies, according to some surveys. "As long as everything fits on the screen and it's bright, they can't tell the difference." Please don't let this get in your way or convince you that the screen is 'obviously' what it says on the tin in light of such evidence. Read up on PenTile (which this reminds me of), take Darkstone's test (pay attention to color mismatch and pixelation!) and so on.

 

*shrug*

added my own 1080p screen after I bought it on the l702x... tis 100% 1080p

 

SXPS 16 here: I see the rings at any resolution, any resizing.

Couldn't the lack of rings just mean that the other device is bluring everything together?

 

I have an L502X w/ 1080p screen. I see rings on the face in that picture both when zoomed and not zoomed.

 

I see the rings:
- with my XPS L702x internal monitor (1920x1080 and 1600x900)
- with my XPS L702x with an external monitor (HDMI@1920x1200)
- with my desktop (1920x1200)

 

The screens are definitely 1080p what's wrong is that they are still TN panels which is why all the quality isn't as good as one might think, compared to higher quality of IPS panels.

 

There appears to be some confusion in regards to the type of rings I'm talking about -- there are gradient rings (which are normal), and then there are the concentric, high-relief rings forming distinct colored bands at the boundaries of the normal gradations as produced by the dithering. I'm talking about color distortions that are quite different at native resolution than at any other zoom level or resolution.

Try this test; it's a bit better: Clock and phase - Lagom LCD test

If the huge square is pixelated and different in color (or shade of gray) from the reference image, that's a big problem. For me, the square appears light beige/yellow and pixelated at native resolution and zoom. Well, there is no light yellow/beige color in that test page, I checked; this is evidence that the monitor is using some kind of hack to 'fake' the resolution at 1080p.

 

Don't know what it means but mine looks yellow at 1920*1080 and grey at all other resolutions in Chrome/Firefox/Opera.

But on Internet Explorer the same image its GREY at 1920*1080!

So, now what? Confused...


Edit:

Ok, Internet Explorer as some sort of bug that reports the resolution to be 1536*864.

 

First of all I want to say I've hated Pentile screens since I first saw them, exactly because of the funky color fringing.
Having said that, I really dont think the screen on my xps is pentile, all tests that I can do lead me to think it's just e mediocre 1080p screen.
About the clock/phase test: I looked really close at the screen and I can see it is indeed a mesh of white and black pixels, just as it should be (here's what the test says it should look like: The test image below is best viewed in full-screen mode and should appear grey from a distance, but from close by, you may notic that it is a fine pattern of interleaved black and white pixels)
IMO if the monitor were pentile you would not see black and white pixels in succession but a weird fringing phenomenon due to the lack of subpixels. As it is, if you look closely you can see all the pixels, but from a distance on my monitor looks yellowish, due to the AG coating, bad colors of TN, etc).
Also bear in mind that the clock/phase test should have little to no effect on screens that are connected via a digital cable (as ours is). It was a test meant to correct the errors from ADC/DAC conversions on LCD screens connected via an analog connection. This test should also display patterns, like rectangles, squares etc on any resolution except the native resolution.
Correct me if I'm wrong but I can't find any evidence this screen is not 1080p (though I clearly agree it's not the best screen, it dithers rather badly, has weird color temperature, etc).

 

My screen is beige on native res and grey at any other resolution be it on any 4 of my L702X 1080p laptops. However if I plug into my Samsung 1080p it shows a grey.

 

I tried this on the l502x 1080p. The image appears fine at 1080p, though I do see a little bit of banding (small bands though) on the test on that same site. However, the effects are greatly heightened when I switch to 1600x900. At that res, the grey image looks blue and there is definite banding. So, does that imply that this issue is isolated to the 702?

 

That is not a bug. That means your DPI setting is at 125% (check this!) or you're zoomed in by about 130%.

 

I have DPI and zoom in browser at 100%.

I just reported what was written in this page, when i tested with IE: Display settings - Lagom LCD test

But i probably had zoom at +100% like you say, because i tested again and now that page reports the right resolution.

 

I think we're talking color dithering, not resolution differences. Because if you actually use a flat color and go pixel by pixel, you will see the whole screen fill up one at a time, and the pixel boundaries are clear.

Often cheaper LCD panels on laptops use 6-bits per color and use dithering to make up the remaining 2-bits, which as far as I know can result in the behavior you're seeing.

EDIT: Nevermind, missed the second page.

 

The clock and phase test is not only something you should view at a distance but is also extremely sensitive to gamma shift and color tinting. Did you profile your monitor? Are you using Firefox's color management?

I'm not sure how the banding that you're seeing isn't just a consequence of static dithering on your 6-bit panel along with the ordered dithering on your picture. Have you tried viewing the original source on a true 8-bit panel or using error correction dithering?

 

It appears properly on a true 8-bit monitor with any kind of dithering, but only appears correctly on this monitor with error correction dithering. The fact that it doesn't appear correctly here is the issue -- a series of two similarly colored regions of different dither ordering shouldn't produce a bright line at their border.