Dell Alienware AW3423DW 34″ QD-OLED 175Hz (3440 x 1440)

Late to the party. This seems to be the only monitor to get if picture quality is your main concern. There are no other QD OLED monitors that I am aware of. There are some with LG panels, but the colors would be lacking in comparison.
Picture quality of the very few IPS based panels like the X27, etc is better as they have pixel density over this, and they are substantially brighter which is nice to have for brighter HDR scenes. However, I generally prefer this panel for gaming due to the true black of OLED, and the fact that you're going down to 1440 versus 4k which helps frame rates.

If picture quality was my main concern and I was doing professional type work I would not get this monitor. However, I do agree it's the best currently for high hz HDR gaming.
 
Picture quality of the very few IPS based panels like the X27, etc is better as they have pixel density over this, and they are substantially brighter which is nice to have for brighter HDR scenes. However, I generally prefer this panel for gaming due to the true black of OLED, and the fact that you're going down to 1440 versus 4k which helps frame rates.

If picture quality was my main concern and I was doing professional type work I would not get this monitor. However, I do agree it's the best currently for high hz HDR gaming.
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
 
  • Like
Reactions: Xar
like this
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
Depends on the content.

I have said many times - This panel is far better for gaming than my X27. However, the X27 I still keep plugged in for all other productivity work.

There is a reason Apple isn't using OLED for their stand-alone displays. It's micro-led IPS, and believe me - No one doing professional work is using OLED right now. You'd be an idiot to do so.
 
I've had my AW3423DW for like 2 or 3 months and LOVE it, the OLED vibrant colors cannot be beat, and deep blacks are literally BLACK not dark grey, and smooth fluid motion of the image. Tried the LG C2 42 and it's too large, and the colors were not as vibrant.
 
Depends on the content.

I have said many times - This panel is far better for gaming than my X27. However, the X27 I still keep plugged in for all other productivity work.

There is a reason Apple isn't using OLED for their stand-alone displays. It's micro-led IPS, and believe me - No one doing professional work is using OLED right now. You'd be an idiot to do so.
I think gaming is assumed to be the primary use here.
 
I was a day 1 adopter. Better brighter picture than my b9 55" and more smooth. I haven't had the same fan sound issue people are talking about, but text definitely has that rainbow effect. Tried that 3rd party "Bettertext" or whatever and it seems to help some, but still very noticeable. Not a huge deal since I'm 95% gaming on it most of the time.
 
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
QD-OLED and WRGB-OLED w/ MLA are way above any Mini-LED and Edge-Lit LCDs in IQ.

2024 Phosphorescent Blue QD-OLED will take lead as the best panel technology until Micro-LED fully established itself in the consumer market.
 
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
I always found VA panels to look the worst.
 
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
That's way too reductive. Each of these technologies has its tradeoffs.

IPS tends to do well for desktop use as it has decent response time, viewing angles, color accuracy etc but low contrast ratio, IPS glow...
VA has higher contrast ratio but bad viewing angles andmore issues with pixel response times for dark transitions. Samsung has mostly solved the pixel response times in their models.
OLED has per pixel local dimming, excellent viewing angles and response times, but low peak/sustained brightness, pixel structures that don't work well for desktop use and potential for burn in.

Throw mini-LED backlights into the LCDs and results are varied - higher peak/sustained brightness than OLED but often at the expense of pixel response times and blooming due to not having per pixel dimming.

You just pick which compromises are acceptable to you. I won't pay the massive prices for e.g PG32UQX (mini-LED IPS) because its pixel response times suck for how expensive it is, I'm not looking to buy the Samsung Neo G7/G8 because of the issues those models have, I am not interested in QD-OLED because I need a 70% desktop use, 30% gaming display so the pixel structure issues are too much and the resolution is too low. It would be fine as a pure gaming display but I can just use my OLED TV for that.

Personally I don't care too much what technology is used but rather look at these things on a per display model basis. In the past about 8-10 years or something I've gone 30" 2560x1600 60 Hz IPS -> 27" 2560x1440 144 Hz TN -> 49" 5120x1440 120 Hz VA -> 48" 3840x2160 120 Hz OLED -> 28" 3840x2160 144 Hz IPS and am now looking to get a Samsung 57" 7680x2160 240 Hz VA, depending on final specs and pricing.
 
That's way too reductive. Each of these technologies has its tradeoffs.

IPS tends to do well for desktop use as it has decent response time, viewing angles, color accuracy etc but low contrast ratio, IPS glow...
VA has higher contrast ratio but bad viewing angles andmore issues with pixel response times for dark transitions. Samsung has mostly solved the pixel response times in their models.
OLED has per pixel local dimming, excellent viewing angles and response times, but low peak/sustained brightness, pixel structures that don't work well for desktop use and potential for burn in.

Throw mini-LED backlights into the LCDs and results are varied - higher peak/sustained brightness than OLED but often at the expense of pixel response times and blooming due to not having per pixel dimming.

You just pick which compromises are acceptable to you. I won't pay the massive prices for e.g PG32UQX (mini-LED IPS) because its pixel response times suck for how expensive it is, I'm not looking to buy the Samsung Neo G7/G8 because of the issues those models have, I am not interested in QD-OLED because I need a 70% desktop use, 30% gaming display so the pixel structure issues are too much and the resolution is too low. It would be fine as a pure gaming display but I can just use my OLED TV for that.

Personally I don't care too much what technology is used but rather look at these things on a per display model basis. In the past about 8-10 years or something I've gone 30" 2560x1600 60 Hz IPS -> 27" 2560x1440 144 Hz TN -> 49" 5120x1440 120 Hz VA -> 48" 3840x2160 120 Hz OLED -> 28" 3840x2160 144 Hz IPS and am now looking to get a Samsung 57" 7680x2160 240 Hz VA, depending on final specs and pricing.

The 57" is interesting for desktop use, but not sure it will be possible to buy it anytime soon. Seems like a niche low-yield product that they'll be in no hurry to start selling.
 
IPS tends to do well for desktop use as it has decent response time, viewing angles, color accuracy etc but low contrast ratio, IPS glow...
VA has higher contrast ratio but bad viewing angles andmore issues with pixel response times for dark transitions. Samsung has mostly solved the pixel response times in their models.
OLED has per pixel local dimming, excellent viewing angles and response times, but low peak/sustained brightness, pixel structures that don't work well for desktop use and potential for burn in.
Something I will say about IPS and its low contrast for desktop use that it isn't always a bad thing. High contrast can strain the eyes if you go at it for a long time. Actual paper has a fairly low contrast ratio, maybe 250:1 or so in the case of good paper and inks, less in the case of newsprint or the like. Yet people often will prefer a physical book because it is easier on the eyes compared to a screen. The reason is partially less contrast, partially the fact that since it is reflective the amount of light matches the surrounding light.

Also 100% doesn't matter in an office, the room lights are so bright that the reflections from the monitor ruin your contrast ratio.
 
I was planning an upgrade to this display but am ditching my 1900R for an LG C2 42" as my PC gaming display.... and a 4080 for native hdmi 2.1 and 4k grunt ... I like the ultrawide for FPS games but not RPG... missing that full face 4k real estate for those. That said the 1900R has been a great display for the last few years.
 
Just hit a year of mixed WFH and gaming with this monitor. No signs of image retention which is good but I've been pretty careful so it would have been pretty disappointing if there was any.

Aside from the dumb self power off issue it's been great and I doubt I will upgrade until high refresh 4k ultrawide oleds are a thing
 
Just picked up on dell outlet for $870 (780+ 90 shipping, tax)

only 90 days warranty tho, this should hold me over for high hz gaming till 4k 144hz+ becomes an affordable reality.
 
Possible OLED "ClearType" Breakthrough!

Founder of Blur Busters / TestUFO here.

As I am working very hard on a "ClearType for OLEDs" for all OLEDs (both LG WOLED and Samsung QDOLED) via a third party font renderer (MacType for Windows)

The third party font renderer software at mactype.net is customizable with horizontal/verticals in subpixel rendering, so in theory could be made compatible with Samsung QD-OLED triangular-subpixel structure; some early tests are promising.

For ClearType lovers who "upgraded" to an a QD-OLED, please test this...

Beta Test Image Eliminating Green/Purple Fringing

If you have a Dell Alienware AW3423DWF or another QD-OLED, please test this image:



Download for viewing in image viewer (to avoid zoomed browser): custom-triangle-cleartype-for-QDOLED-displays.png

Please specify if 1, 2, 1b, 2b, 1c, 2c, 1d, 2d has the least green/purple fringing on your QD-OLED display. Once I know the result, I'll communicate to the author of MacType for some OLED profile configuration files.

I'm going to obtain some prototype LG WOLED MacType profiles too.

UPDATE: Someone else independently came up with a fix on reddit 9 months ago:
ACTUAL fix for the AW3423DW sub-pixel layout/text fringing.
Still working on LG WOLED verison.
 
Last edited:
The 2's have far less or no fringing to my eyes. 2,2b, or 2d are the best. I can't notice much difference between them. Everything on the left has obvious fringing.
 
Out of curiousity, I've Airplayed this image to my 65" 4k Sony A95K QD-OLED and it all looks like a trainwreck to me lol. Insignificant improvements in text fringing, so might as well go with the default.
 
BTW does anyone have "Dark Mode" text clarity examples as well?
I see that is being ignored most of the time, even though many of us use Dark Mode(s) almost exclusively. I've seen it myself on this AW model and it's even worse then black text on white. Text fringing is so horrific in Dark Mode I've given up on QD-OLED for monitor in approximately 1 second.
Apparently the LG OLED pixel layout is better for Dark Mode, but the information on this is sadly very much lacking. It looks like no one buys these OLEDs for mixed used, just for gaming (which is absolutely glorious on these OLEDs, of course).
 
BTW does anyone have "Dark Mode" text clarity examples as well?
I see that is being ignored most of the time, even though many of us use Dark Mode(s) almost exclusively. I've seen it myself on this AW model and it's even worse then black text on white. Text fringing is so horrific in Dark Mode I've given up on QD-OLED for monitor in approximately 1 second.
Apparently the LG OLED pixel layout is better for Dark Mode, but the information on this is sadly very much lacking. It looks like no one buys these OLEDs for mixed used, just for gaming (which is absolutely glorious on these OLEDs, of course).
I use Dark Mode in almost all situations on WOLED (LG 48CX, 42C2), including Windows UI, websites, and UWP apps. I am very sensitive to text fringing and don't notice fringing at all during routine use. I use the TVs as my main PC monitor (work, study, and gaming). When I get close and really scrutinize text, the only areas I see fringing are in certain areas of Windows UI. I don't see fringing in UWP apps or Chromium. I feel that text could be sharper on the WOLED, though.

I should mention, though, that I have to disable ClearType in order to not see any fringing.
 
Last edited:
Really should clarify what effective PPD you are using when reporting such things imo. It's not about how it looks under a microscope, it's about how it looks at normal viewing angles. And at normal viewing angles, a 42" 4k or 48" 4k screen (any 4k screen at the human 60 deg to 50 deg central field of vision) is going to have a much higher PPD than a ~ 27inch-like 13.x inch tall 1440h one.

You have to go by the screen surface to your eyeballs though, not the desk dimensions if you want to be precise. There can be some variance there too depending on what you are doing - if you are over your desk with your peripherals on top of it or if you are more casually laid back using a game controller in your chair for example.



Pixel sizes in regard to text on the 2D desktop and in regard to 3d game engine graphics are more or less compensated for at 60PPD but only because text sub sampling and aggressive AA in games are applied to mask how large the pixel structure actually is. As most people know, unfortunately LG OLED uses WRGB and samsung uses pentile which are both non-standard subpixel layouts which text-ss is not designed for. The 2D desktop's graphics and imagery typically have no pixel size/edge-masking compensations at all so are displayed at the "raw" pixel grid/size and the true granularity. So even higher than 60 PPD is better all around. The smaller the perceived pixel sizes, the less noticeable artifacts and fringing issues are. In text and even occasional edge artifacts from things like DLSS and frame insertion/amplification technologies. Larger perceived pixel sizes, larger problems. (and vice-versa)

. . . .

The pasted comment below are mostly in regard to 42" 4k and 48" 4k screens. A 3440x1440 . . . (440w +[ 2560w x1440] + 440w) . . when viewed at "full height" to your perspective - like a 2560x1440 would be at 60deg to 50 deg viewing angle - would only be around 43 PPD to 51 PPD. You could sit farther away than that but the height would shrink to your perspective making the screen look even shorter and more belt-like.

So if someone makes a perfectly mapped for wrgb or pentile text sub-sampling it would be improved compared to text-ss on non standard subpixel layouts as it is now of course but - you still won't have great edges on 1400 - 1500p like pixel sizes to your perspective (1440p screens or 42" 4k screens at ~ 24" or so view distance ). You'd just be back to what text looks like on rgb subpixel layouts at ~ 43 PPD to 51 PPD which isn't great compared to 64ppd - 70 PPD - 77PPD you'd get on a 4k screen within a 60 to 50 deg horizontal viewing angle. Properly mapped text-ss would allow 60PPD+ range on 4k screens at 60 to 50 deg viewing angle to look great on text though, where now you probably have to use 70PPD+ to shrink the actual pixels sizes small enough (and either use 3rd party text subsampling, greyscale, etc. or disable text-ss entirely). As mentioned, text-ss in general doesn't do anything for aliasing on the 2d destkop's graphics and imagery so higher PPD is better anyway if you can get it.
. . . . .

https://qasimk.io/screen-ppd/

At the human viewing angle of 60 to 50 degrees, every 4k screen of any size gets around 64 to 77 PPD.

At the human viewing angle of 60 to 50 degrees, every 2560x1440 screen of any size gets only 43 PPD to 51 PPD.

At the human viewing angle of 60 to 50 degrees, every 1920x1080 screen of any size gets only 20 PPD to 25 PPD


. . . . . .


>This graphic shows the optimal viewing distances of both a 42" 4k and a 48" 4k, plus a lower 50-ish PPD one at 24" to show how a lot of people are viewing one on a desk sub-optimally at \~ 1500p like PPD.
>
>
>
>
842627_tJWvzHy.png


\-.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-

>Human viewing angle is 50 to 60 degrees
>
>
3kU3adt.png

>
>optimal viewing angle minimizes the off axis and non-uniform edges
>
>
XvKRu9t.png

>
>
>
>
>
>Sitting too close pushes the sides of the screen outside of your viewpoint and makes the off-axis areas larger:
>
>
>
>
RUdpoK8.png




. . . . . . . . . . . .

TLDR: When relaying how text aliasing/fringing, graphics aliasing, etc. look on different monitor sizes and specs it's more meaningful if you include the PPD (or the true viewinng distance to your eyeballs at least). A 42" 4k at ~ 24" view on a desk is only ~ 51 PPD which maps pretty close to what ~1500p would look like on a desktop monitor. Someone viewing text at 64PPD or ~70PPD with optimal viewing angles on a 4k that is decoupled from the desk is going to see thing much differently - seeing much smaller pereceived pixel sizes, much less granularity.
 
Last edited:
Really should clarify what effective PPD you are using when reporting such things imo. It's not about how it looks under a microscope, it's about how it looks at normal viewing angles. And at normal viewing angles, a 42" 4k or 48" 4k screen (any 4k screen at the human 60 deg to 50 deg central field of vision) is going to have a much higher PPD than a ~ 27inch-like 13.x inch tall 1440h one.

You have to go by the screen surface to your eyeballs though, not the desk dimensions if you want to be precise. There can be some variance there too depending on what you are doing - if you are over your desk with your peripherals on top of it or if you are more casually laid back using a game controller in your chair for example.



Pixel sizes in regard to text on the 2D desktop and in regard to 3d game engine graphics are more or less compensated for at 60PPD but only because text sub sampling and aggressive AA in games are applied to mask how large the pixel structure actually is. As most people know, unfortunately LG OLED uses WRGB and samsung uses pentile which are both non-standard subpixel layouts which text-ss is not designed for. The 2D desktop's graphics and imagery typically have no pixel size/edge-masking compensations at all so are displayed at the "raw" pixel grid/size and the true granularity. So even higher than 60 PPD is better all around. The smaller the perceived pixel sizes, the less noticeable artifacts and fringing issues are. In text and even occasional edge artifacts from things like DLSS and frame insertion/amplification technologies. Larger perceived pixel sizes, larger problems. (and vice-versa)

. . . .

The pasted comment below are mostly in regard to 42" 4k and 48" 4k screens. A 3440x1440 . . . (440w +[ 2560w x1440] + 440w) . . when viewed at "full height" to your perspective - like a 2560x1440 would be at 60deg to 50 deg viewing angle - would only be around 43 PPD to 51 PPD. You could sit farther away than that but the height would shrink to your perspective making the screen look even shorter and more belt-like.

So if someone makes a perfectly mapped for wrgb or pentile text sub-sampling it would be improved compared to text-ss on non standard subpixel layouts as it is now of course but - you still won't have great edges on 1400 - 1500p like pixel sizes to your perspective (1440p screens or 42" 4k screens at ~ 24" or so view distance ). You'd just be back to what text looks like on rgb subpixel layouts at ~ 43 PPD to 51 PPD which isn't great compared to 64ppd - 70 PPD - 77PPD you'd get on a 4k screen within a 60 to 50 deg horizontal viewing angle. Properly mapped text-ss would allow 60PPD+ range on 4k screens at 60 to 50 deg viewing angle to look great on text though, where now you probably have to use 70PPD+ to shrink the actual pixels sizes small enough (and either use 3rd party text subsampling, greyscale, etc. or disable text-ss entirely). As mentioned, text-ss in general doesn't do anything for aliasing on the 2d destkop's graphics and imagery so higher PPD is better anyway if you can get it.
. . . . .

https://qasimk.io/screen-ppd/

At the human viewing angle of 60 to 50 degrees, every 4k screen of any size gets around 64 to 77 PPD.

At the human viewing angle of 60 to 50 degrees, every 2560x1440 screen of any size gets only 43 PPD to 51 PPD.

At the human viewing angle of 60 to 50 degrees, every 1920x1080 screen of any size gets only 20 PPD to 25 PPD


. . . . . .


>This graphic shows the optimal viewing distances of both a 42" 4k and a 48" 4k, plus a lower 50-ish PPD one at 24" to show how a lot of people are viewing one on a desk sub-optimally at \~ 1500p like PPD.
>
>
>
>

\-.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-.-.--.-

>Human viewing angle is 50 to 60 degrees
>
>
>
>optimal viewing angle minimizes the off axis and non-uniform edges
>
>
>
>
>
>
>
>Sitting too close pushes the sides of the screen outside of your viewpoint and makes the off-axis areas larger:
>
>
>
>

​

. . . . . . . . . . . .

TLDR: When relaying how text aliasing/fringing, graphics aliasing, etc. look on different monitor sizes and specs it's more meaningful if you include the PPD (or the true viewinng distance to your eyeballs at least). A 42" 4k at ~ 24" view on a desk is only ~ 51 PPD which maps pretty close to what ~1500p would look like on a desktop monitor. Someone viewing text at 64PPD or ~70PPD with optimal viewing angles on a 4k that is decoupled from the desk is going to see thing much differently - seeing much smaller pereceived pixel sizes, much less granularity.

So what are the optimal viewing distances for say a 27" 16:9, 34" 21:9, and 42" 16:9?
 
Out of curiousity, I've Airplayed this image to my 65" 4k Sony A95K QD-OLED and it all looks like a trainwreck to me lol. Insignificant improvements in text fringing, so might as well go with the default.
IMPORTANT: Cannot use AirPlay for the Demo Images
  • AirPlay isn't always 1:1 native pixel mapping; it often downrezzes
  • In some cases, AirPlay automatically uses 4:2:2 or 4:2:0 chroma depending on device, which is known for years to destroy ClearType rendering
  • Compression will diminish/erase the subpixel rendering.
Unless you can configure AirPlay to do native 4K 4:4:4 chroma at better-than 4K Blu-Ray bitrates, the best-practice is to always disable subpixel rendering during streaming/casting operations anyway. Subpixel rendering is chiefly designed for uncompressed video cable use; e.g. HDMI or DisplayPort, at native resolution with overscan disabled (Aspect Ratio configured to "PC" or "Just Scan" or "1:1" setting).

...

Also:

Via non-AirPlay, the LG OLED demo image has extremely sharp text

With the LG OLED demonstration image having nearly twice as sharp text than any of the other options ...

(viewing the LG OLED images on anything else other than LG OLED, will not work -- it is as sharp as well-optimized LCD ClearType).

There are ClearType fans that would like the sharpest possible text. Some of are very picky.

Everybody sees differently.

  • Different degrees of eyevision acuity and eyeglasses prescriptions.
  • Different degrees of colorblindness (12% of population is colorblind).
  • Different people are bothered by different things more than others.
  • Some of us don't notice the color fringing as blatantly as the text-fuzziness
  • Others like the zero-subpixel MacType look better.
  • This also applies to motion too (stutter vs tearing vs blur vs etc)

It is a matter of personal preference

 
Last edited:
So what are the optimal viewing distances for say a 27" 16:9, 34" 21:9, and 42" 16:9?


Human central view is 60 to 50 degrees.


There is some wiggle room there but that is optimal, focal + color field. Past that you are pushing the sides of the screen well out into your periphery (as well as lowering your PPD more, making the pixel sizes appear larger).
60 deg 4k is 64 PPD but the 60PPD point isn't pushing the screen edges out by that much more at 65deg viewing angle or so. Non standard subpixel layouts and the uncompensated for 2d desktop's graphics and imagery can benefit from a lot higher than 60 PPD but it's a good reference point for rgb text with rgb sub-sampling (or any text-ss mapped appropriately to other non-standard subpixel structures in the future).

That one link I showed is pretty useful. You can plug in any size screen and just scale the distance up/down until you see 50 deg and 60 deg.

https://qasimk.io/screen-ppd/

At the human viewing angle of 60 to 50 degrees, every 4k screen of any size gets around 64 to 77 PPD.

At the human viewing angle of 60 to 50 degrees, every 2560x1440 screen of any size gets only 43 PPD to 51 PPD.

At the human viewing angle of 60 to 50 degrees, every 1920x1080 screen of any size gets only 20 PPD to 25 PPD



Ultrawides, super-ultrawides, and curve radii complicate things a bit.

If you look at an ultrawide as a central 16:9 + wings on the sides, you might consider that central 16:9 monitor space as where you'd want to get near to the 60 to 50 degree viewing angle, using the rest of the screen for peripheral immerison (driving, flying games especially), and as head/eye turning desktop app real estate like a multi-monitor setup but without bezels. You could sit farther away than that and get somewhat higher PPD as well as seeing more of the screen's width in your central view all at once, and some people do, but then the screen height would shrink to your perspective so it isn't optimal to sit much farther vs perceived screen height after a point. Imo then it ends up like a short belt instead of a "full height" field.

With curved screens the R-value is the radius of the screen, which equates to the focal point if the curved screen was considered a lens. E.g. 1000R = 1000mm = around 40" radius/focal point. Most people end up having to sit too close to a curved screen so that the focal point of the curve is a location farther behind them unfortunately. That means the pixels outside of the center of the screen won't be pointed at your eyes. Another way to say that is that they are off axis.

As an analogy, think of them as a grid of laser pointers on the curve - in a room with a fog machine. If you were sitting at the focal point of a large 800R (~35" radius/focal point) curved screen for example, every point on the screen would be pointed directly at your eyeballs so you'd be seeing just the points of light almost entirely. If you were sitting say 26" away from a 1000R curved screen that was mounted on a desk, all of the laser pointers would be pointing at a location 14" farther behind you (40" radius). In the fog machine room in the 26" desk view distance example you'd be seeing oblong points and the sides of the light shafts instead of the lights pointing directly at you on the sides, and to greater degree/longer the farther away on the curve. That results in more image distortion like a funhouse mirror. People blame game devs for not "fixing" this with warping compensation methods but if you had a curvature more appropriate to your screen size vs view distance this would probably be much less of an issue in some games. Some devs do literally stretch the displayed content though like tv's "scale 4:3 = stretch" modes did and that is a problem.

. . . .

To answer your 34" 21:9 optimal view distance/optimal viewing angle question, I'd say that depends on what game you are playing and whether you are using it for desktop/app real-estate at any given time, and probably your personal taste to a degree.

A 34" 3440x1400 screen could be looked at essentially as a 27" 16:9 2560x1440 with 440px of wings on each side, bezel free.

440 + [27" 2560] + 440

In order to get the full height to your perspective you'd have to stay near to the 60 to 50 degree central viewing angle on the middle portion. You could choose to essentially shorten the screen to make it more belt like and see more of the sides of the screen in your central viewing angle though, fitting the entire screen width closer to 60 to 50 degrees. That would be more like cutting a screen in half vertically rather than expanding a full height screen's width if you get what I mean. 1/2 field height + full field width . . vs . . full field height + wider than full field.

2560 x 1440 27" 16:9 screen 60 degrees and 43 PPD at ~ 20.5" view distance <--- same for 34" 3440x1440 ultrawide if using full field height + sides as peripheral

2560x1440 27" 16:9 screen 50 degrees and 51 PPD at ~ 25" view distance <--- same for 34" 3440x1440 ultrawide if using full field height + sides as peripheral

.

3440 x 1440 21:9 screen 60 degrees and 57 PPD at ~ 27" view distance <--- "half" field height, full field width

3440 x 1440 21:9 screen 50 degrees and 69 PPD at ~ 34" view distance <--- "half" field height, full field width



The 3440x1440 alienware is 1800R, , 1800mm . . 71" radius (or focal point) which is extremely broad, a very slight curve section of a very large circle so I probably wouldn't even take the curve into consideration in this case with this small of a screen.

. . .

You can sit closer of course but when you sit closer than the focal point of the curve the curve distorts things since the pixels are instead pointing at a location behind you rather than where you are sitting.

The solid blue is the 1000R viewpoint, the transparent field is when sitting nearer. There is some wiggle room like the first example but the 2nd example is poor imo.


0UhdIIr.png




q03mqmG.png


The 42" and 48" 16:9 optimal viewing angles and what distances are in the other image I posted (along with a 24" view distance I added to show a sub-optimal desktop placement)

. .
 
Last edited:
Human central view is 60 to 50 degrees.


There is some wiggle room there but that is optimal, focal + color field. Past that you are pushing the sides of the screen well out into your periphery (as well as lowering your PPD more, making the pixel sizes appear larger).
60 deg 4k is 64 PPD but the 60PPD point isn't pushing the screen edges out by that much more at 65deg viewing angle or so. Non standard subpixel layouts and the uncompensated for 2d desktop's graphics and imagery can benefit from a lot higher than 60 PPD but it's a good reference point for rgb text with rgb sub-sampling (or any text-ss mapped appropriately to other non-standard subpixel structures in the future).

That one link I showed is pretty useful. You can plug in any size screen and just scale the distance up/down until you see 50 deg and 60 deg.





Ultrawides, super-ultrawides, and curve radii complicate things a bit.

If you look at an ultrawide as a central 16:9 + wings on the sides, you might consider that central 16:9 monitor space as where you'd want to get near to the 60 to 50 degree viewing angle, using the rest of the screen for peripheral immerison (driving, flying games especially), and as head/eye turning desktop app real estate like a multi-monitor setup but without bezels. You could sit farther away than that and get somewhat higher PPD as well as seeing more of the screen's width in your central view all at once, and some people do, but then the screen height would shrink to your perspective so it isn't optimal to sit much farther vs perceived screen height after a point. Imo then it ends up like a short belt instead of a "full height" field.

With curved screens the R-value is the radius of the screen, which equates to the focal point if the curved screen was considered a lens. E.g. 1000R = 1000mm = around 40" radius/focal point. Most people end up having to sit too close to a curved screen so that the focal point of the curve is a location farther behind them unfortunately. That means the pixels outside of the center of the screen won't be pointed at your eyes. Another way to say that is that they are off axis.

As an analogy, think of them as a grid of laser pointers on the curve - in a room with a fog machine. If you were sitting at the focal point of a large 800R (~35" radius/focal point) curved screen for example, every point on the screen would be pointed directly at your eyeballs so you'd be seeing just the points of light almost entirely. If you were sitting say 26" away from a 1000R curved screen that was mounted on a desk, all of the laser pointers would be pointing at a location 14" farther behind you (40" radius). In the fog machine room in the 26" desk view distance example you'd be seeing oblong points and the sides of the light shafts instead of the lights pointing directly at you on the sides, and to greater degree/longer the farther away on the curve. That results in more image distortion like a funhouse mirror. People blame game devs for not "fixing" this with warping compensation methods but if you had a curvature more appropriate to your screen size vs view distance this would probably be much less of an issue in some games. Some devs do literally stretch the displayed content though like tv's "scale 4:3 = stretch" modes did and that is a problem.

. . . .

To answer your 34" 21:9 optimal view distance/optimal viewing angle question, I'd say that depends on what game you are playing and whether you are using it for desktop/app real-estate at any given time, and probably your personal taste to a degree.

A 34" 3440x1400 screen could be looked at essentially as a 27" 16:9 2560x1440 with 440px of wings on each side, bezel free.

440 + [27" 2560] + 440

In order to get the full height to your perspective you'd have to stay near to the 60 to 50 degree central viewing angle on the middle portion. You could choose to essentially shorten the screen to make it more belt like and see more of the sides of the screen in your central viewing angle though, fitting the entire screen width closer to 60 to 50 degrees. That would be more like cutting a screen in half vertically rather than expanding a full height screen's width if you get what I mean. 1/2 field height + full field width . . vs . . full field height + wider than full field.

2560 x 1440 27" 16:9 screen 60 degrees and 43 PPD at ~ 20.5" view distance <--- same for 34" 3440x1440 ultrawide if using full field height + sides as peripheral

2560x1440 27" 16:9 screen 50 degrees and 51 PPD at ~ 25" view distance <--- same for 34" 3440x1440 ultrawide if using full field height + sides as peripheral

.

3440 x 1440 21:9 screen 60 degrees and 57 PPD at ~ 27" view distance <--- "half" field height, full field width

3440 x 1440 21:9 screen 50 degrees and 69 PPD at ~ 34" view distance <--- "half" field height, full field width



The 3440x1440 alienware is 1800R, , 1800mm . . 71" radius (or focal point) which is extremely broad, a very slight curve section of a very large circle so I probably wouldn't even take the curve into consideration in this case with this small of a screen.

. . .




The 42" and 48" 16:9 optimal viewing angles and what distances are in the other image I posted (along with a 24" view distance I added to show a sub-optimal desktop placement)




. .

Thank you for taking the time for this.
 
Anyone else having issues with this monitor pixel refresh not completing in 7 mins and the entire time it's slowly blinking Red as I write this it's still going at minute 23 and counting.
 
...

Everybody sees differently.

OK, since you took the time to explain everything so nicely, for the first time, I hooked up my MacBook M1 Pro 16" over HDMI to my Sony A95K 65" 4k QD-OLED.
I guess Mac OS could make a big difference or even break your test, but "for science", I've selected 4k resolution, no scaling, full screen, and all the examples still looked bad. Of course, when I scaled to 1080p, everything looked so much better. All in all, I got drawn the most to QD-OLED 1d and 2d, but I am thinking this is all extremely subjective, since the differences are so small, so I wonder if all this is even worth the effort.

Just for kicks, I've tested a coding IDE (IntelliJ Idea) with 1440p and 1080p scaling. It looks wonderful, and I can't see anyone complaining about text fringing or text clarity on QD-OLED and I have a wonderful, razor sharp MacBook M1 Pro IPS FALD with massive amount of dimming zones to compare it with.
I also noticed, that Dark Mode (white letters on dark background) now somehow looks better than Light Mode (regular black letters on white) on 4k QD-OLED (with scaling). I was convinced it was the other way around on QD-OLED since I saw the AW 1440p QD-OLED myself and it was a mess in Dark Mode...
 
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.
For brightly lit room on good IPS I would say picture quality is not terrible and rather completely adequate.
Not great like on OLED or even plasma but adequate. Black level in this scenario doesn't bother me because black still looks black.
Similarly generally bright content doesn't require very low black level. Image will still look better with black being actually black just like eg. bright content looks better on IPS than CRT (which has very low ANSI contrast ratio) but it doesn't matter as much.
Also for normal desktop usage, especially using normal black on white color schemes IPS is absolutely ok.
When comparing IPS to OLED it was obvious IPS doesn't display true black... still it looked adequate even with direct comparison.

It is much worse with pitch black room and because of that I rarely didn't use ambient light when using IPS panels.
Still I would not say IPS is terrible because this is ...

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
... how I would describe all VA panels

At first glance VA might look better than IPS, especially in dark room but image on VA always looks off and no amount of ambient light can make it look correct because VA just displays colors incorrectly and each eye sees different ratio of colors.
 
OK, since you took the time to explain everything so nicely, for the first time, I hooked up my MacBook M1 Pro 16" over HDMI to my Sony A95K 65" 4k QD-OLED.
I guess Mac OS could make a big difference or even break your test, but "for science", I've selected 4k resolution, no scaling, full screen, and all the examples still looked bad. Of course, when I scaled to 1080p, everything looked so much better. All in all, I got drawn the most to QD-OLED 1d and 2d, but I am thinking this is all extremely subjective, since the differences are so small, so I wonder if all this is even worth the effort.

Just for kicks, I've tested a coding IDE (IntelliJ Idea) with 1440p and 1080p scaling. It looks wonderful, and I can't see anyone complaining about text fringing or text clarity on QD-OLED and I have a wonderful, razor sharp MacBook M1 Pro IPS FALD with massive amount of dimming zones to compare it with.
I also noticed, that Dark Mode (white letters on dark background) now somehow looks better than Light Mode (regular black letters on white) on 4k QD-OLED (with scaling). I was convinced it was the other way around on QD-OLED since I saw the AW 1440p QD-OLED myself and it was a mess in Dark Mode...

For more science, curious what your view distance is on the 65" 4k ? The perceived pixel size makes a huge difference in how much stepping/fringing/aliasing.

Someday we'll have resolutions so high we won't have to mask how blocky the appear at viewing distance in the first place. 8k will probably have to lean on masking of the pixel edges less breaking 4 pixel blocks per each one of 4k but we'll still need to mask the edges some for a while yet.
 
For more science, curious what your view distance is on the 65" 4k ? The perceived pixel size makes a huge difference in how much stepping/fringing/aliasing.
Most of the time, I was pixel peeping from like 1 ft (cca 30cm), but I did try to look from a couple of feet away too, as far as I could go, and actually read the text comfortably. IDK, the sample still looked bad to me personally (without any scaling applied). But with scaling (at least 1440p) and in Dark Mode, I am now convinced, you could 100% work with 4k OLED every day. I use scaling on non-OLED 32" as well, so it's something I always use for nicer experience.

BTW for reference of a sort, I recently returned the "famed" VA 4k 32" Samsung Neo G7, because it was so annoying to work on, in spite of very sharp text. In Dark Mode, the bottom of the screen would be washed out in a really distracting way and that already legendary curve is just unbelievably distracting in best case, to outright unusable for work in worst case. Nothing like that is happening on the 4k QD-OLED at all. Now they just need to make a 32" 4k version of it...
 
Thanks for the info.

Scaling adds more pixels per character in a font similar to how using a higher resolution does so isn't a great measure though it is a usable workaround a lot of people use for example when getting ~ 1500p like PPD of around 51 - 53 with a 42" OLED on a desk (at the cost of losing full 1:1 4k worth of desktop real-estate down to more like 1500p worth). Otherwise the non-standard subpixel layout would have alot worse fringing with fewer pixel blocks to define each font with. Scaling doesn't do anything for the unmasked 2d desktop's graphics and imagery for the most part though, and you can't scale games vs aliasing either (outside of some interface scaling some games have to a degree) as it doesn't work the same way for game graphic objects that remain the same size in the scene relative to the screen size.

Distance changes the staircase/fringing/grid size to your eyes so your view distance vs the ppi (measured as PixelsPerDegree) is the most important factor to begin with. Staring up close under a microscope isn't that meaningful, even magazine print is made up of tiny dots. It's the end result that is important.


. . . .
. . . .


pixel.zoom.doom.gif


. . . .

. . . .


pixel.zoom.game.art.gif



. . .


anti aliasing sort of smudges out pixels to hide how blocky the pixels sizes actually are to you at a given distance. Better types affect edges more, and DLSS has it's own sharpening and AA . . but they all (AA and text-ss) mask part of the images, at least on contrasted edges, to blur out the fringing/blockiness in one way or another of the actual poor size of pixel grids

pixel.anti.aliasing_baby.toon_1.gif



. . .

DLSS will look better the higher the resolution it is starting from (you can't get blood from a rock), and any edge artifacts from DLSS and/or frame insertion will be less visible the tinier the pixels look to your eyes in the first place (just like fringing from text and non-standard subpixel layouts vs text-ss types will be less visible). Eventually when we move to 8k screens we'll probably start upscaling 4k to 8k instead of 1080p->1440p -> 4k too so will get even cleaner results.

The higher the frame rate, the less difference between frames so the better things like dlss and frame amplification will work as well. Smaller is better. Smaller perceived pixel sizes due to ppi vs view distance (PPD) is better, smaller periods and differences between new frames (higher fpsHz) is better.

wiki:

TAA compared to DLSS​

Nvidia's DLSS operates on similar principles to TAA. Like TAA, it uses information from past frames to produce the current frame. Unlike TAA, DLSS does not sample every pixel in every frame. Instead, it samples different pixels in different frames and uses pixels sampled in past frames to fill in the unsampled pixels in the current frame. DLSS uses machine learning to combine samples in the current frame and past frames, and it can be thought of as an advanced TAA implementation.

. .

https://www.hardwaretimes.com/taa-vs-dlss-is-nvidia-dlss-better/

combines temporal upscaling with NVIDIA’s proprietary neural net upscaler. The latter compares high-quality 16K images against the base resolution, learning the upscaling of the latter into the former. DLSS 2 works on the same basic principle as TAA, leveraging temporal feedback to increase the number of samples per frame. However, unlike TAA, it doesn’t sample every frame’s pixels.

Much like checkerboard rendering, DLSS samples different pixels in different frames, using the temporal data from previous frames to fill in the gaps. The ever-learning neural network blends the present pixel with the previous ones.

. . .

https://www.rockpapershotgun.com/dlss-3-how-it-works-how-it-performs-and-when-you-should-use-it

. . . . . . . . .

I wonder if we'd be able to use AA / DLSS's AA on the desktop if we rendered the desktop on the side of a 3D cube for example? Wonder how that would work out.

. . . . . . . . . .


, because it was so annoying to work on, in spite of very sharp text. In Dark Mode, the bottom of the screen would be washed out in a really distracting way and that already legendary curve is just unbelievably distracting in best case, to outright unusable for work in worst case.

That's a VA screen so any off-axis pixels are going to be non-uniform with the middle of the screen. Some VA screens worse than others.

Curved screen's R value is the radius of the curve. Most screen's size vs the focal point of that curve result in the viewer sitting with the focal point well behind them which will cause distortion as the farther away the pixels are to the sides of the screen, the more off axis they are to you, pointing past you so you are looking at them more and more degrees towards sideways. The G7 has a 1000R curve on a screen that is only 32" 16:9 and meant to sit on a desk. 1000R = 1000mm = around 40" radius or focal point of the curve. The pixels are all pointing at a place probably 16" behind you when viewing a 32" screen at 60 degrees ~ 24" view distance so it will be distorted and off axis the further from the center. The solid blue is where every pixel on the screen surface is pointed directly at the viewer horizontally.



840386_q03mqmG.png
 
Last edited:
Yes, but the text clarity/fringing is not an issue for gaming, just for productivity, and productivity i.e. staring at a lot of text all day long, is what scaling makes a much nicer experience of.

If you'll allow me just a small off-topic note about the Neo G7 1000R curve...It doesn't look like a perfect curve in person. It's like it starts very deep in the middle and then flattens out (i.e. moves away from you) on the sides. It's extremely weird and distracting. At least that's the subjective experience.
 
IPS panels have terrible picture quality. I have a highly rated IPS gaming monitor and the picture quality is terrible compared to my OLED television.

If you look at a place like rings, IPS panels rank below VA panels, which rank below OLED panels, with QD OLED panels being the best.
Little late to this discussion but I just wanted to chime in and say that IPS are still the gold standard for color grading professionals. Not all IPS are created equally of course, but the best monitors for accurate color grading appear to still be using IPS panels.
 
Most of the time, I was pixel peeping from like 1 ft (cca 30cm), but I did try to look from a couple of feet away too, as far as I could go, and actually read the text comfortably. IDK, the sample still looked bad to me personally (without any scaling applied). But with scaling (at least 1440p) and in Dark Mode, I am now convinced, you could 100% work with 4k OLED every day. I use scaling on non-OLED 32" as well, so it's something I always use for nicer experience.

BTW for reference of a sort, I recently returned the "famed" VA 4k 32" Samsung Neo G7, because it was so annoying to work on, in spite of very sharp text. In Dark Mode, the bottom of the screen would be washed out in a really distracting way and that already legendary curve is just unbelievably distracting in best case, to outright unusable for work in worst case. Nothing like that is happening on the 4k QD-OLED at all. Now they just need to make a 32" 4k version of it...
There's very annoying interacting scaling settings -- where 100% DPI needs to be scaled to work properly. The Amazon FireTV, I believe, for example applies a scaling setting in some software (e.g. browsers) that you have to "undo"; and the Aspect setting may interfere with another TV menu setting. Much like how you get 100% in Chrome on a 200% DPI display by going 50% zoom in Chrome (it nets out 100%). So sometimes scaling is needed to get the subpixels to click correctly, somehow. Also there's an accessibiltiy scaling setting (forgot where), where if set to 80%, you need 125% DPI in order to get exactly 100%. It's a confusing soup, sadly.

UPDATE: Someone else independently came up with a fix on reddit 9 months ago:
ACTUAL fix for the AW3423DW sub-pixel layout/text fringing.
However, I'm still trailblazing for a fix for LG WOLED.
 
Last edited:
Yes, but the text clarity/fringing is not an issue for gaming, just for productivity, and productivity i.e. staring at a lot of text all day long, is what scaling makes a much nicer experience of.

If you'll allow me just a small off-topic note about the Neo G7 1000R curve...It doesn't look like a perfect curve in person. It's like it starts very deep in the middle and then flattens out (i.e. moves away from you) on the sides. It's extremely weird and distracting. At least that's the subjective experience.

Interesting about the curve... Some monitors are flat in the middle with curve on the wings instead.

regarding text - some pc rpg games have very fine text in stat blocks and trees, inventory/item properites, even text chat boxes and npc text etc. Consoles tend to have more jumbo versions for obvious reasons but some pc rpgs and rts games have a lot of small text and of different colors. There is usually some flexibility with interface sizing in modern games but then you are crowding your screen more and/or seeing less info at a time.

====================================================

3D game engine AA isn't available on the desktop , only text AA on fonts, so the 2D desktop's graphics and imagery get no AA typically at all.

I've wondered if MS would ever make a 3D mode for the desktop, mapping it to a face of a 3D cube so AA / DLSS AA could be active on the desktop.

Apps like this one make a snapshot of the desktop mapped to a 3d cube. You can't actually use the desktop in the 3d mode, it's just a static texture used for rotating the cube as a transition.

http://www.otakusoftware.com/deskspace/

screen_4.jpg


There was an app called bumptop that made a 3d windows environment but it was bought up by google for use on android and so was discontinued.

As far as I know there have been no apps that allow the 2d desktops actual running app windows to be in 3d space, or even rendered flat onto a front facing cube (rather than rotating axis object capability) in a 3d engine where graphics anti aliasing could be active.

. . .

I did find something interesting in looking that stuff up:

Windows mixed reality works with VR headsets but I thought it would be neat if you could use the same tech to run a 1:1 pixel resolution in 3D as your windows desktop locally without having a VR headset plugged in. Then I found some info that you could enable develper mode and try it out. Could be interesting if MS ever made a full resolution 3d desktop for actual desktop use with some full screen anti aliasing options on it.

https://learn.microsoft.com/en-us/w...pts/using-the-windows-mixed-reality-simulator

r0XMZsK.png
 
Last edited:
Back
Top