We need to go to 8k (I see inbetween pixels).

[Wayne Morellini]

Being feeling better lately, and went into a rare Vmax session. I was shocked because I noticed I could see a screen door effect between the pixels at certain times, had to really look for it though. I thought it must have been a 2k screen, now I suspect that maybe it is the off sub pixels. Seems to be something below what your brain can zoom in on in individual pixels, but still can see it as a patch of pattern if you look for it, so you don't notice it. Ran into the manager today and found out it is supposed to be a 4k screen. Standing in front of a LG 84inch today, I could see the same thing.

4k is yesterday's technology, we need 8k at the top end. The capability of doing 4k has been around for many years. Early last decade we had 4k lcd monitors, and I seem to remember a crt in the 80's. It is so cheap to make that around 2007 NVIDIA wanted to make it a new monitor standard to fit a new graphic mode, and manufacturers were talking about introducing 4k TV. Now Ultra HD 4k TV's are nearly $1k in the States. It is not a top end performance technology to justify large price tags. We should be going 8k in pro equipment, such as $50k cameras.. 8k bayer footage downgraded should also make for better 4k pictures. It also has usage in various large outdoor advertising, photo editing use and other computer use and surround entertainment, and computer table use, and conferencing.

Already Sony is planing 16k+ holographic cameras, so 8k should not be thought of as the end of the line and held back.


4k is dead as a pro only product.

 

[Brandon J.F.]

Yeah, there are already 4k capable camera sensors for phones.: http://www.androidauthority.com/aptina-4k-video-image-sensors-163538/

And 4k tablets: http://connect.dpreview.com/post/0766715919/panasonic-4k-photographers

 

[Wayne Morellini]

Thanks Brandon. Soon their is going to be 4k glasses, and more.

 

[Gavin Greenwalt]

If you see between pixels that's a function of the pitch more than the resolution. For instance I have an Occulus VR headset and the biggest problem is the dot pitch. It's a 1280x720 screen but looks like way lower resolution due to FOV and because so much of the screen is actually black. A good OLPF in such instances would go a long way.

 

[Hrvoje Simic]

What Gavin said. What you notice is the grid between pixels. The perception mostly depends on pixel pitch, then on the type of pixel arrangement.

Digital projection is the sufficient example, where pixels don't have to be visible even with 2K resolution.

 

[Björn Benckert]

Try do the VFX heavy films that goes around in 8k. There is more than cameras and screening in the equation. Also if you look at a nice made 2k DCP on a 8 meter wide screen sitting about the middle of the theater. I can only say that I do not see pixels. I know 4k looks better but still "the man on the street" does not care much about it, and they actually decide becuase the producers listen to them when they buy ticket and when they do not. And so far I do not think there is a film that has been making bank just because it was mastered in 4k. Especially not if you take into the consideration what the film could have been if it was mastered in 2k. In 2k you get a few CG dinosaurs for the same price as you get some wire removals in 4k.... so I do not think 8k mastering will come any day soon. 4k will come but there will be a lot made in 2k still.

 

[Wayne Morellini]

If you see between pixels that's a function of the pitch more than the resolution. For instance I have an Occulus VR headset and the biggest problem is the dot pitch. It's a 1280x720 screen but looks like way lower resolution due to FOV and because so much of the screen is actually black. A good OLPF in such instances would go a long way.

Gavin and Hrvoje,

The black lines actually have a finer resolution than the pixels, so you actually are seeing more resolution. Fov was acceptable width. A olpf (a strange way of putting it) just hides this greater resolution, even lowering sharpness.


What people don't understand, is that olpf is not raising resolution, but washing it out, lowering contrast, and lower the resolution of the dot of the pixel in the pitch. It is both biological and a trick that this makes a resolution better, it is a compromise. Firstly, as you pool out the focus of light it modem's with surrounding light, diluting accuracy. As contrast between pixels lowers at the same time they get closer to to each other in level, which means it becomes harder to see the resolution than for a black pixel with white neighbouring pixel. So no, it does not make the resolution better, inter pixel space is not the real issue here. Ironically, what is actually happening with this scheme, is that the pooling is creating circular gradients of much higher resolution, and higher resolution features. As the light of a pool falls away it makes a continuous tonal gradient, that our eyes Max out on in their resolution ussually (depends on number of photons and size). These overlaps between pool of light also result in higher Res features smaller than a pixel. The circles themselves are firmed analogly mostly, so are another extra high resolution feature. You are actually introducing beyond 8k information to hide the low resolution.

In old CRT TV's, the pixels were hidden by blooming (your eyes do a little too). Similar to the olpf idea, they, and the beam, spread their energy a cross to their neighbors. If you take the same picture and display it on a crt and a modern LCD, at the same size, reasonably good brightness, you will probably be amazed how bad the video frame looks on the LCD (unless the LCD is using uprezzing techniques or smoothing, or you need to check your eyesight). The particular resolution features you are missing in the square pixels compared to circular/maybe pools, is the hard almost triangular corners. If you sat down and did calculations of the size relative to a circle, you can work out how much finer resolution The bulk of the corner is, probably around 2k using 720*540 SD content, as the thin corners of the triangle in the non bulk corners go so fine resolution, they are beyond our eyes ability to see it. If the pixel was nearer the limit of our vision, we should not see sharp corners on it, eventually with resolution, the pixel becoming more like a dot.

 

[Wayne Morellini]

Another affect of this contrast reduction, that makes it harder to see the difference between pixels, say the brightness level resolution, is that people enjoy contrast. So even though you might not see a pixel, you enjoy the resulting increased in contrast in details from the increased contrast in the pixels. A particular problem with Bayer olpf and glossy compression. Other issues are, vision impairment in focus or neuron performance that makes things fuzzy, and not looking with the right part of the eye (the small spots of 2400 dpi vision the eye scans a scene with) which might not be dead center or people have gotten off it, but you can train yourself to look for it with fine detail patterns (which is what I was using). But where ever somebody notices it or not, there is a segment of the population that sees it, and 8k is a good compromise for the vast majority, and ones who still can look to see probably won't notice too much (as with me with the 8k plus inter pixel spaces).