Demostrating Pimax 8k has real RGB subpixels with an 4k panel resolution. And explaining the source of its image problems


#42

I am suspect that how much advantage if pimax have a 4k native 60Hz mode for 8k, ever have someone asked for this and they tell that they will check the possibility.

Unfortunately they don’t tell us before making the decision. Switching to 5k+ will lose this chance?

@xunshu @PM_Sean


#43

This is what the sub pixels look like to me:


#44

Yeah that’s what I see and is what I’m talking about above.


#45

Okay folks let’s over simplify this & go 1 2 1.

  1. Grid perpendicular to the horizon (5k) can draw perfect lines North to South & East to West

  2. The 8k grid is 45° to the horizon (diagonal) it can draw perfect lines NE to SW & NW to SE.

So for the letter “R” to look as good as in no.1 on no.2 the letter needs to be rotated 45°.

This example is without scaler or sub pixels. The simple truth the 8k cannot be as clear as the 5k due to pixel grid being 45° opposed.

No using zoomed pictures just plain facts.


#46

Well, we are not really talking about image clarity and quality. We are talking about the pixel and subpixel layout and count. The results are devastating.


#47

We are talking about 50% of pixels calculated by graphics card are wasted in panels problems and added to the disposition of customers significantly increases the SDE to the point of leaving it almost equal to 5k +, if this is the sums that humans are basing their constructions and productions in a horizontal and vertical concept that can not be seen clearly in these headset I think the 8k is not what I expected. I think that is not fair what was promised.

Even so, pimax is close to creating a good headset, pimax solved bigger problems than this with the lenses, I think what they need is the change of panel, Pimax knows all of about the panels probles, this is why we have the 5k + and being less can update for free, but the backers of the 8k, I think that pimax can not afford it, that’s what I say to share the costs or that we make a proposal that we can study the extra cost and those who would pay it.

I would be willing to pay for it and even wait a bit.


#48

I’m a little uncertain how a link to an article explaining how a different display technology was turned askew to hide the fact that they were faking their resolution supports the claim that this panel would have full colour per pixel. I have also played the role of trying to match the evidence to a full pixel count; it ends in embarrassment.

Also note that while these images are higher resolution than earlier ones (revealing, for instance, the clear vertical RGB pattern in the 5K+ panel), it’s still through the VR headset lens. That means massive CA and some notable diffusion. The relative placement of the colours varies a lot across even these closeup images. What doesn’t vary is the fact that even for green, traditionally the highest resolution channel, this panel is simply missing half the pixels.


#49

What about Heart Shaped pixels.?


#50

Outstanding. Thank you. :slight_smile:


#51

I wonder if any entities are trying to make a deep learning upscaler.


#52

Sure, we can pick that orientation. Did you want your pixels double height or width? It’s equally applicable either way, and either way you’re left with a sub-spec resolution on an axis. The relative position of the colours in the region shown is mostly CA.


#53

@Heliosurge yes that seems to be a symptom of using a diagonal pixel arrangement.
Why have the panel makers used a diagonal pixel arrangement is what you should be asking.
A pixel needs to have red/green/blue to count as a pixel.
If you define the pixel horizontally then the vertical adds up correct but the horizontal is halved.
If you define the pixel vertically then the horizontal pixels add up but vertical is halved.
Just look at the pictures. Define the pixels either way & count.
This panel has a diagonal pixel arrangement in order to immitate a 4k panel while using less pixels.


#54

As far as I can tell however there aren’t any other appropriate sized 4k displays out there. I had a poke and I found exactly one. Guess what it advertised it’s pixel pattern as? “Rainbow-RGB” Now I’ve never seen that description before but it sounds remarkably like a euphemism for what we are seeing.

Pimax have used this panel because it’s all there is.


#55

That’s what DLSS is in Nvidia’s demos. 1440p upscaled to 4K.


#56

Damn. I really had hoped the issue was panel utilization and that Pimax could fix it by using better lenses, but I can’t deny the evidence in this thread. This means the 8k is a dead-end and even worse, the Pimax 8k-X is too!! Damn …


#57

Why if there using Hisense(formerly sharp) is not a good thing.

Sharp was decent & hi sense is not. Even lg got dinged with 4k no so panels on TVs.

Someone mentioned this type often used in projectors?


#58

The 4k was rainbow rgb.


#59

Yep. You can count them either way. The thing is The pixels are asymmetrical. But each pixels have its RGB sub-pixels. The panel has all 8.3M sub-pixels for each color. Not pentile sharing green. I would not call them sub-spec resolution. I would say it is Alt-spec resolution because it maintains its pixel density. Dont’ get me wrong, clarity suffers a lot for not being squared like all images it have to render that have square pixels requiring re-sampling to maintain aspect ratio.


#60

Well remember the Sharp panel in the 4k is called rainbow rgb as well.


#61

Ahh interesting. In that case that must refer to the offset diagonal pixels. Hmm in that case the 8k panel is not showing up so my parts suppliers can’t access this low persistence CLPL manufacturer. Oh well, might be hope then.