The lines between the pixels

[yuriythebest]

Greetings! Some users have reported that the space between the pixels can be a bit distracting- is it Possible to hide it by adjusting the lense so that stuff is more blurry?

[nateight]

This is the "screen door" effect. Short answer: Stuff is already going to be way too blurry without additional filters.

Longer answers:

"The resolution can be partially fixed by using a diffusion filter to blur the edges between the pixels, it works very well. Problem is, it hurts contrast, brightness, and black levels pretty heavily."

You may also be interested in the Reddit discussion here, and Michael Abrash's GDC talk here. The ultimate conclusion you're likely to reach is that there is simply no single magic bullet, all we can do is support VR in its imperfect present state and wait for the technology to catch up to our outsized expectations.

[WiredEarp]

I'd imagine a custom diffusion layer could be made. It would have to be aligned, but would basically have 'holes' to let the centre of the pixel show brightly, while diffusing the pizel border zone.
Would not be as perfectly even as diffusing the entire screen, but would still reduce the 'screen doors' without such a strong loss of brightness and clarity.

[geekmaster]

In my opinion, the "screen" effect is like wearing a "bee bonnet" on your helmet. You are looking through a screen attached to your head. This is especially apparent in the Unity Tuscany demo with the transparent text overlay, which looks like it is "printed" on a spherical screen at about 18-inches from your face. Just think of it as protection from insects or other small projectiles, in place of a perspex face mask...

Once you accept it as part of your head gear, no problems, mate!

I just took some photos to demonstrate. Here is the original image shot from my computer screen (the vertical lines are aliasing between my LCD pixels and my camera pixels):

host_tuscany.jpg

And here is a closeup shot with my 58mm Canon Close-Up lens held against the eyecup hole in the Rift (after removing one eyecup):

rift-screendoor1.jpg

Click the photo to see it fullscreen. That photo is EXACTLY what I see when wearing my Rift (except that this is a small portion of the full screen in the Rift). As you can see, it looks just like a insect repellant screen over a helmet, so just accept it as head gear, like I said, and it will not break the immersion...

Interestingly, the 58mm lens fits between the eyecup clips like it was meant to be mounted there.

[geekmaster]

I'd imagine a custom diffusion layer could be made. It would have to be aligned, but would basically have 'holes' to let the centre of the pixel show brightly, while diffusing the pizel border zone.
Would not be as perfectly even as diffusing the entire screen, but would still reduce the 'screen doors' without such a strong loss of brightness and clarity.

WTF? How do you diffuse NO LIGHT? With no visible photons being emitted by the area BETWEEN the pixels, there are no photons to "diffuse". Did you study physics?

[MrGreen]

Thanks for the pics, GM. Doesn't look that bad to me.

[Evenios]

another VERY nice video someone did of TF2 video "inside" the rift

[youtube-hd]http://www.youtube.com/watch?v=UG8a0oOlAog[/youtube-hd]

you can clearly see that while a) theres some motion blur and some screendoor effect its not that bad and the poster said watching this video on a 40 inch TV pretty much gives the same effect as viewing it actually in the rift quality wise..... so. To me thats not bad at all first First try VR. and the issue about "blurriness" some people have stated might be they did not have it centered right. theres a sweet spot too so .

a big sigh of re-leaf to me thinking from some "impressions" that it was bad. Decent enough to me and will look better with 3d depth people tend to complain too much i guess (myself included)

[LeeN]

I am near sighted and I can see there is a noticeable difference between wearing glasses and not wearing glasses.

With out glasses using the C lens, I get distracted a lot more by the pixel spacing to the point where it's hard to read text at times, but when I wear my glasses everything seems fine to me and I don't notice the pixel spacing as much. It also may be related to having to increase the lens distance from my eyes to wear glasses.

That said, I can't wait for them to start offering consumer grade development kits!

[Zoide]

I'd imagine a custom diffusion layer could be made. It would have to be aligned, but would basically have 'holes' to let the centre of the pixel show brightly, while diffusing the pizel border zone.
Would not be as perfectly even as diffusing the entire screen, but would still reduce the 'screen doors' without such a strong loss of brightness and clarity.

WTF? How do you diffuse NO LIGHT? With no visible photons being emitted by the area BETWEEN the pixels, there are no photons to "diffuse". Did you study physics?

Come on, there's no need to treat people that way.

[GeraldT]

The original RAW image shows all screendoor lines the same, but in this JPEG image, they seem to have periodic variation, but more "regular" near the displayed text, which must be caused by a JPEG artifact.

It a moire effect caused by scaling the picture, you get them also from projectors with their screendoors when using certain types of screens.
Please upload a non-scaled image, that would be really helpful to see how it really looks. Thank you.

[Likay]

Greetings! Some users have reported that the space between the pixels can be a bit distracting- is it Possible to hide it by adjusting the lense so that stuff is more blurry?

That's exactly how i adjust the focus of my projectors. Sharpest focus does not (in my opinion i have to add) give the best image. However a very slight focusloss setting is prefered to not mess up the contrast too much. The screendoor effect is still visible but not as disturbing as with a sharp focus. It's up to each and everyone to find the best setting but it's definitely worth the effort to experiment with.
I guess the golden setting also depends on the resolution used. My projectors are old ones with 1024x768 pixels which may explain my prefered focus settings.

With an hmd where the eyelens also is an active part of the focusprocess i doubt it will work though. I'm not entirely sure that a diffuser will really give a noticeable lift either but my own experience with such solutions is nonexistent.

[OzOnE2k10]

@geekmaster - I think @WiredEarp was trying to say that the diffuser would obviously "borrow" some of the light from adjacent (LIT) pixels to help fill in the gaps.

Maybe it is possible to make a diffuser which "magnifies" each pixel by a tiny amount so the resulting pixels overlap the gaps?
Then again, I'm sure many manufacturers have tried such things in the past, and it doesn't seemed to have worked out too well.

So far (judging from the pics / vids), the OR Dev Kit looks great - the screendoor is far less than I'd imagined.

I agree with your "bee bonnet" comment. The S3D and headtracking likely more than makes up for a LOT of these minor issues.

I just need my damned Hydis panel to arrive. (I've only tested @Zacherynuks's panel before).
I'm sure we could hugely reduce any motion blur by doing a straightforward backlight strobing mod.

Does anyone fancy trying it on a Hydis?
I should be able to knock together some simple Arduino / AVR code for testing this.

It would require finding a VSYNC signal on the controller board though.
Apparently, the RTD2662 chip used on many controllers can output VSYNC on pin 97 - does anyone have an o'scope to test this? (it may be disabled in the firmware though)

OzOnE.

[Parallaxis]

I don't quite get this obsession about screen door and low res. Once you start accepting this, you can enjoy the VR experience and start making some content.

The resolution and screen door effect will get better whether you guys are whining about it or not.

The blur I can understand, because it seems to be related to a whole bunch of factors like how fast your GPU is, software, drivers, how tight your HMD is to your face and so on.

[OzOnE2k10]

The resolution and screen door effect will get better whether you guys are whining about it or not.

Yep, I'm SO happy to be getting something way better than my crappy old VGA "video" goggles with the postage stamp sized image.

I really don't think the screendoor thing will matter too much at this point - the Dev kit is already a HUGE leap forwards for VR.

Many high-PPI panels are coming within the next six months! Sit tight.
The consumer version is gonna be mindblowing.

[WiredEarp]

WTF? How do you diffuse NO LIGHT? With no visible photons being emitted by the area BETWEEN the pixels, there are no photons to "diffuse". Did you study physics?

Oh, grow up geekmaster. Even if I was a total noob making a dumb mistake, your comment would still be unacceptable. I'm sure you mean well, but you should restrain your inner Doc Brown a little. Fortunately, people aren't required to have any sorts of degree to do good things with VR, or to contribute to these forums.

In this case, and as OzOnE2k10 kindly pointed out, I was of course referring to having a hole in the centre of each pixel element, so that the majority of the light comes through undiffused. I didn't think that was too difficult a concept for most people to understand...

@ OzOnE2k10, your ideas about using a magnifying screen I thought were interesting. Don't some glasses free lenticular systems already basically use sheets of micro lenses? However, as everyone else has pointed out (and I agree), theres not much point in investing much effort into avoiding SDE. Your mind will just accept it after a while, and technology will advance quickly and make it a moot point. However, its always good to discuss the possibilities for its reduction.

[OzOnE2k10]

Yeah, I'm guessing the lenticular sheets can be made extremely precisely nowadays.
Even more so with the crazy 3D printing tech that's already out there.

When you can 3D print a spaceship the width of a human hair, you can imagine the lenses they could now make...
[youtube]http://www.youtube.com/watch?v=wThtfAtB5U8[/youtube]

As you say, I'm not too worried about SDE atm.
The Rift is like going from 56K dial-up to Broadband (trust me, we were one of the last streets in town to get ADSL. We even resorted to ISDN and satellite internet before that).

btw, I just contacted njytouch on eB@y to ask if they thought a panel controller could be made to strobe the backlight directly in firmware.
Hopefully (if they understood my message OK), this could lead to a firmware upgrade for the Rift to reduce any panel-related blurring.

Their controller programming kits are cheap atm, so could be fun to play with. (I won't link directly here).
I don't know if they have Realtek programming cables, but I'm hoping so.

[geekmaster]

Come on, there's no need to treat people that way.

Okay, sorry. I thought he was suggesting to diffuse darkness by scattering missing photons...

The overabundance of speculative misinformation around here is astounding, and getting me more than a little annoyed. Things that exist only in Virtual Reality are fine, but Virtual Facts are not fine unless they also exist in the Real World. There is a thing called Google, which people can check there facts before saying something that could get them laughed at in Physics class...

In this case, more details were provided in a later post explaining that only part of the outer edges of the pixels would be diffused in his proposal. That cannot work well if you look at how how the pixels contain RGB subpixels, so on the sides you would only be diffusing a gradient from blue to red across the blackness, and the center hole in the diffusion layer would show mostly green. Again, physics, reality, facts, Google...

[anon84]

another VERY nice video someone did of TF2 video "inside" the rift

[youtube-hd]http://www.youtube.com/watch?v=UG8a0oOlAog[/youtube-hd]

I posted something like this on oculusrift.com but the forum is quite slow, maybe I can get an answer here, at the end of the vid 3:10 on, the screendoor effect is completely removed for very brief moments when his camera is adjusting focus I believe (it's hard but you can try and pause the vid on these exact frames to see it better) there is no loss of contrast or brightness as has been suggested would be the result of using a filter

Is there some way to replicate the 'out of focussness' and use the rift like that all the time?
it would be like a FSAA effect with 0 screendoor, personally I prefer a bit of blur in my games anyway to add realism, nothing in life is razor sharp unless you are very close to it or have eyes like a hawk

just thought, a question for rift owners, they mention a 'sweetspot' in the rift, can you just purposely avoid the sweetspot to blue the image a little or is that impractical due to the tiny adjustments required, only for a slight movement to mess it up again.

[geekmaster]

another VERY nice video someone did of TF2 video "inside" the rift ...

I posted something like this on oculusrift.com but the forum is quite slow, maybe I can get an answer here, at the end of the vid 3:10 on, the screendoor effect is completely removed for very brief moments when his camera is adjusting focus I believe (it's hard but you can try and pause the vid on these exact frames to see it better) there is no loss of contrast or brightness as has been suggested would be the result of using a filter

Is there some way to replicate the 'out of focussness' and use the rift like that all the time?
it would be like a FSAA effect with 0 screendoor, personally I prefer a bit of blur in my games anyway to add realism, nothing in life is razor sharp unless you are very close to it or have eyes like a hawk

Trying to photograph through the lenses does not work well unless the camera lens models your eye lens and is held the same distance from the center of the Rift lens. I provided photos earlier in this thread that were taken through the eyecup hole WITHOUT the eyecup, using a macro closeup lens stack in my SLR camera. They show what the screendoor really looks like to the naked eye through the Rift lenses (especially around the transparent text areas where there are less JPEG artifacts).

No amount of software processing will make the screendoor effect go away (or even change). It is a hardware thing. You cannot smear photons onto the black areas with software. You could use a diffusion layer (such as rubbing a little Vaseline petroleum jelly onto your Rift lenses), but that harms image quality in addition to reducing screendoor effect (which is already vastly superior to older LCD technology).

With lenses that have adjustable focus, the recommended way to fix screendoor is to defocus just enough to blur the pixels together so they are touching but not overlapping. A similar method is used when calibrating CRTs, where you electrically defocus the beam until adacent scanlines touch, even though that sacrifices horizontal resolution in closed-circuit applications.

[anon84]

yes I know software can not fix it.

looked at your shot, looks the same as in the vid, the blur in your image is real enough for me though, wouldn't want to add any more than that

what about the motion blur, can you really see the last 3 frames of your reticle like that

[KBK]

I'd imagine a custom diffusion layer could be made. It would have to be aligned, but would basically have 'holes' to let the centre of the pixel show brightly, while diffusing the pizel border zone.
Would not be as perfectly even as diffusing the entire screen, but would still reduce the 'screen doors' without such a strong loss of brightness and clarity.

WTF? How do you diffuse NO LIGHT? With no visible photons being emitted by the area BETWEEN the pixels, there are no photons to "diffuse". Did you study physics?

Black body radiation?

Interesting how humans think there is a state of nothing and somehow...it can affect everything. Human perception says 'it is a nothing', reality says 'interference-regardless'

(pay no attention to the insanity behind the curtain - of human perception)

It's soooo much easier to poke you, Geek. Geek the angry science guy.

But seriously, diffusion screens to not work in these applications. Unless you publish in mid 70's Penthouse, that is.

[geekmaster]

Black body radiation? ... But seriously, diffusion screens to not work in these applications. Unless you publish in mid 70's Penthouse, that is.

According to my thermal imaging cameras, black bodies emit about the same radiation as caucasian bodies (especially when undressed). Of course, I have no color preference, but I prefer viewing female thermal radiation...

A bit of related trivia: Bob Guccione actually smeared Vaseline petroleum jelly on his camera lenses (except in the very center) to get that peripheral diffusion in his Penthouse photos.

Yeah, the rampant speculative misinformation does get under my skin a little, especially AFTER I already published contrary REAL facts on the same subject. Sorry about that. I should just go play with my Rift and let you guys duke it out amongst yourselves, right?

[Tirregius]

I'm about a half a diopter close in my left eye and a bit better in my right...I'm hoping this is just about the right focus correction to help the screen door effect without losing resolve!

Whatever I get, that is what I plan to design for. I think it is a worthwhile step to see how we can make the most out of this early design.

I want to go on record as saying that i believe it is a bit counter-productive to have modded Rifts floating around. Ideally, new content available to our tiny vr community should be experienced on reference hardware. If we are all running different screens at some point, our ability to evaluate software solutions will be hampered.

[geekmaster]

... I want to go on record as saying that i believe it is a bit counter-productive to have modded Rifts floating around. Ideally, new content available to our tiny vr community should be experienced on reference hardware. If we are all running different screens at some point, our ability to evaluate software solutions will be hampered.

Of course, our software needs to support standard Rifts and standard eyecups, but I plan to ALSO support Rift mods with my software.

Even PalmerTech (Palmer Luckey) is a modder. For that matter, without his modding fetish we would not even HAVE the Rift to play with! Thanks Palmer (and his AWE-culus crew)!

[Tirregius]

This is a different phase, Geekmaster, where they are preparing to bring a product to market. I can also see the benefit of having hardware mods out there. I can get my hands dirty myself, and I could easily see myself grabbing a higher density panel to drop in there...but then if I'm running 4x the resolution of what is expected to be out there, how valuable is my interpretation of software soln's? Maybe I wrong and the benefit of diverse hardware outweighs the downside, but my instinct says "let Oculus, I.e. "our modder/hacker captain" Palmer Luckey put the hardware he wants us to test on in our hands. If they come out with an official mod, which I think is possible but not likely, then that makes more sense.

[GeraldT]

This is a different phase, Geekmaster, where they are preparing to bring a product to market. I can also see the benefit of having hardware mods out there. I can get my hands dirty myself, and I could easily see myself grabbing a higher density panel to drop in there...but then if I'm running 4x the resolution of what is expected to be out there, how valuable is my interpretation of software soln's? Maybe I wrong and the benefit of diverse hardware outweighs the downside, but my instinct says "let Oculus, I.e. "our modder/hacker captain" Palmer Luckey put the hardware he wants us to test on in our hands. If they come out with an official mod, which I think is possible but not likely, then that makes more sense.

There are two groups and directions - those that will create content, they need to keep their Rifts workable with the SDK etc. ... and those that care for the tech and want to find ways to improve on the experience through tech, not software. The latter will give valuable information on what they might not have thought of.
Both are important and if somebody finds a way to give us a 1080p display without breaking tracking/SDK compatibility, then it would certainly help tailoring content towards the consumer Rift.

[geekmaster]

There are two groups and directions - those that will create content, they need to keep their Rifts workable with the SDK etc. ... and those that care for the tech and want to find ways to improve on the experience through tech, not software. The latter will give valuable information on what they might not have thought of.
Both are important and if somebody finds a way to give us a 1080p display without breaking tracking/SDK compatibility, then it would certainly help tailoring content towards the consumer Rift.

And apparently a third "systems" group who stand in both hardware and software "camps", including me, where you are attempting to erect a fence between my feet. Ouch!

[geekmaster]

I replaced the "screendoor" closeup photo above with a better one without JPEG artifacts.

Here it is again so you do not need to scroll back to find it:



That photo is EXACTLY what I see when wearing my Rift (except that this extreme closeup is only a small portion of the full screen visible in the Rift). The actual visible "screen" is much finer, and much less annoying than it may appear in the above photo. As you can see, it looks just like a insect repellant screen over a helmet (i.e. "bee bonnet"), so just accept it as head gear, like I said, and it will not break the immersion...

Bee bonnet:



Also, close your eyes while turning with a controller to help prevent VR motion sickness.

[Parallaxis]

Notice how they still enjoy watching the bees through that grid. They even look immersed. Go figure.

[2EyeGuy]

According to my thermal imaging cameras, black bodies emit about the same radiation as caucasian bodies (especially when undressed). Of course, I have no color preference, but I prefer viewing female thermal radiation...

OK, since I like finding the real answers to crazy random off-topic questions made in jest... This study suggests you are wrong:

In this study, participants viewed images and completed self-report emotional measurements while their facial temperature was measured using infrared thermography. Images varied systematically in their capacity to elicit feelings of potency and activity in a 2(potency: high, low) x 2(activity: high, low) factorial design. The emotion manipulations were derived from the International Affective Picture System (IAPS). We analyzed differences in facial temperature throughout the experiment as a function of self-reported race. We found consistent differences in mean facial temperature by race across all images and all facial regions, but no differences in facial temperature change scores based on race. We also found some minor differences in self-reported emotion by race, but no differences that would explain the consistent temperatures differences observed. We believe that the facial temperature differences observed were due to different skin tone emissivities, which would cause lighter or darker skin tones to be measured at different temperatures by the infrared camera. This theory was supported as we found skin tone to be more predictive of mean facial temperature than was self-reported race.

There's a bit too much political correctness in that article though, so you should take its findings with a pinch of salt.

[geekmaster]

According to my thermal imaging cameras, black bodies emit about the same radiation as caucasian bodies (especially when undressed). Of course, I have no color preference, but I prefer viewing female thermal radiation...

OK, since I like finding the real answers to crazy random off-topic questions made in jest... This study suggests you are wrong:

In this study, participants viewed images and completed self-report emotional measurements while their facial temperature was measured using infrared thermography. Images varied systematically in their capacity to elicit feelings of potency and activity in a 2(potency: high, low) x 2(activity: high, low) factorial design. The emotion manipulations were derived from the International Affective Picture System (IAPS). We analyzed differences in facial temperature throughout the experiment as a function of self-reported race. We found consistent differences in mean facial temperature by race across all images and all facial regions, but no differences in facial temperature change scores based on race. We also found some minor differences in self-reported emotion by race, but no differences that would explain the consistent temperatures differences observed. We believe that the facial temperature differences observed were due to different skin tone emissivities, which would cause lighter or darker skin tones to be measured at different temperatures by the infrared camera. This theory was supported as we found skin tone to be more predictive of mean facial temperature than was self-reported race.

There's a bit too much political correctness in that article though, so you should take its findings with a pinch of salt.

Interesting. I could see how even the existence of this study could be judged as politically incorrect no matter how PC they try to be about it.

If viewed in sunlight, dark skin could be warmer just from absorbing more solar energy (just like wearing dark clothing).

My post was mostly in jest (regarding KBK's "black body radiation" comment). I have not used my thermal imaging cameras to view naked females of any race (yet).

Thermal cameras are also being used to detect potentially contagious people in public spaces who may be ill (i.e. have a fever).

[KBK]

OK, since I like finding the real answers to crazy random off-topic questions made in jest....

OK, since I understand this a little bit better now, would you like me to continue giving you fodder?

I gots plenty.

[Moggle69]

[youtube-hd]http://www.youtube.com/watch?v=fWkGHp4J7dc[/youtube-hd]

[Drewbdoo]

Just something I'd like to say that I haven't mentioned before in comments about the screendoor effect on the devkit and how noticeable it is while playing. I have not tried a rift at all and am still awaiting mine, but I can give you some real world experience that I think is a good analogy for this. I used to be a fencer. In the middle of battle, my mask might as well be invisible for me, aaaand its got a pretty bad screendoor effect:





Edit: Granted, the resolution of the "pixels" is friggin' brillant

[Likay]

Actually the fencemask is "invisible" because your eyes are focused at a point way beyond the mask. Luckily for you i'd say. So the mask serves physically as an "iris" which will make the resulting image a little darker. You can see the fence if you try to focus on the actual fence though.
This will not happen when viewing an hmd where the eyes have to be focused at the display (and therefore also will see the screendoor effect). Imo this is not really a problem and certainly something that one will get used to.

[druidsbane]

At $1500 you don't have a consumer market unlike what the video says. 60 degree FOV and only slightly higher resolution than the Oculus isn't exactly competetive. As long as the Rift keeps moving forward and using the latest tech instead of holding back as many of the mainstays in the industry for some reason keep doing they should be fine Very excited about future revisions and love the current version as it is

[WiredEarp]

@ geekmaster

The overabundance of speculative misinformation around here is astounding, and getting me more than a little annoyed. Things that exist only in Virtual Reality are fine, but Virtual Facts are not fine unless they also exist in the Real World. There is a thing called Google

Yep, you need to work on that. There is a reason people come to these forums, to throw ideas around. If that pisses you off, you need to keep that inside, or at least to a polite level - otherwise we will end up with a place like engadget or youtube, with people slinging more insults than ideas. I have forums where I love to chew out idiots and assholes, but this has always been a polite place and I'd like to see that preserved...

thought he was suggesting to diffuse darkness by scattering missing photons

It always pays to think about other interpretations of the same text.

That cannot work well if you look at how how the pixels contain RGB subpixels, so on the sides you would only be diffusing a gradient from blue to red across the blackness, and the center hole in the diffusion layer would show mostly green.

That is a much more helpful comment, and a good point, although lacks a little imagination. For example, you could simply have a hole in the diffusion element over each subpixel, allowing a percentage of the light from each subpixel to shine through undiffused, boosting brightness when compared to a more basic diffusion layer without holes.

@ OzOnE2k10

When you can 3D print a spaceship the width of a human hair, you can imagine the lenses they could now make...

3D printers seem awesome! Apparently you can't make good lenses out of the consumer ones yet, but I can see that one day this might be a very valid method of making lenses in all sorts of unusual shapes that would be very expensive to develop molds for.
Interesting idea re the backlight strobing - sounds similar to the 'LightBoost' feature on 3D Vision 2?

[geekmaster]

That cannot work well if you look at how how the pixels contain RGB subpixels, so on the sides you would only be diffusing a gradient from blue to red across the blackness, and the center hole in the diffusion layer would show mostly green.

That is a much more helpful comment, and a good point, although lacks a little imagination. For example, you could simply have a hole in the diffusion element over each subpixel, allowing a percentage of the light from each subpixel to shine through undiffused, boosting brightness when compared to a more basic diffusion layer without holes.

Perhaps diffusion can be done VERTICALLY (horizontal stripes only, over horizontal lines and a little of the top and bottom of the pixels). That way the colors should be evenly distributed. But any light that spreads into the vertical black lines would also smear into the colored areas, probably being nearly as harmful to image quality as a solid diffusion layer.

These problems will be solved with more resolution, or with different lenses with more FoV (less magnification). The original prototypes let you see full vertical FoV, and masked off some of the horizontal FoV (with tape, or the USC printable eyecups have a built-in inner mask edge, to be used with wider FoV lenses).

The lenses and eyecups should be cheap and easy to replace, so we can reduce screendoor by using smaller pixels (and more of them visible, depending on Rift adjustments such as eyecups/IPD/screen distance).

[Libertine]

Is the brightness control a backlight adjuster in the Rift? I had thought you might be able to reduce *slightly* the visibility of the "screen door" by raising the backlight.

[realyst2k]

Here's an impractically complex but feasible solution:

Two layers: one LCD, one OLED. The OLED has finer pixels, but fewer of them, in a grid patterns overlaid directly above the LCD display. The grid lightens and darkens in response to the underlying pixel. Since OLED panels are, by default, transparent, it would not occlude the pixel and serve only to alter the in between spaces. Setting it slightly out of focus would cause the color to 'bleed' into the pixel itself.

Of course, this would be rather expensive and impractical unless you are a panel printing titan of likes of Samsung with whatever magical fabs they use. But I like to speculate