Here’s an interesting experiment that lets you simulate PenTile displays on a normal LCD monitor. [Barrett Blackwood] wanted to test out how some graphics look on PenTile RGBG displays with different pixel densities. These PenTile RGBG matrices are sometimes used in OLED displays. For instance, the Nexus One smart phone features a display of this type. Because red, green, and blue OLEDs emit different intensities of light, the pixels are laid out differently from LCD panels in order to balance the color mixing. Our eyes see the green light very well, and so green sub-pixels are made much smaller than their red and blue counterparts.
Because the hardware layout is different, some graphics appear to have crosshatching artifacts in them when viewed on the PenTile displays. [Barrett] made the example above to simulate how graphics look on a traditional LCD screen (image on the left), and how they appear on the PenTile scren (image on the right). The magenta hue seen above is a result of resizing the image. Since the simulation method turns off 1/3 of the green pixels in the image, resizing it ruins the careful calculation. It must be view at a 1:1 ratio to see the image correctly, at which point the magenta magically disappears.
That’s pretty neat.
It’s also interesting that nomatter what the scale, the image appears red on my CRT (which I still use and love).
Also, the same image can be dragged from the CRT (primary) to my LCD (secondary) and it changes color as it goes across, from red to white. The other image doesn’t change at all (as I have my monitors approximately color-matched).
That could be your monitor not being calibrated properly. If the data path from the image to the monitor doesn’t compensate for the monitor’s non-linearity (or the image’s) then two 50% bright pixels won’t equal one 100% bright, which is what this depends on.
This won’t work on CRT’s unfortunately because CRT displays don’t have sub-pixles like LCD’s do. The trick is based on how LCD displays render color.
I have an old LCD Acer X221W display, and 1:1 it looks fine on the center, little purple on the upper and lower side, but when I move my head up and down, the OK looking part moves along.
So, it looks OK only looking straight. Weird image :D
do you per chance have it connected by VGA ?
you are running that lcd in non native resolution, or using analog connection and lcd is not synced properly
You probably have a TN panel. They have really narrow viewing angles. The color shift your describing is visible with any image however any crosshatching will make it much more obvious.
DVI interface, X (Linux) and native resolution. No calibration. Yes, narrow viewing angles. In normal use I don’t see the angle effect, but in that image it’s obvious.
You want to really see an image that will bring your viewing angles to light then check out this one: http://www.lagom.nl/lcd-test/viewing_angle.php
On an IPS display the image will look the same from every viewing angle and the text lagom is barely visible.
I’ve noticed a similar effect when scanning halftoned images in very high resolution. In certain scales the dots blend together to make a normal looking image, but other scales make a moire pattern.
I had a Nexus One with an OLED Pentile display, but I never noticed crosshatching, weird colors, or anything of the sort.
Part of this is due to the simulation. What is not said is that the pentile display also has a different pixel pitch. By faking a pentile display on an LCD monitor you end up with a pixel pitch that is many times greater than that of your phone.
I concur and also have a pentile OLED screen and the display on my phone does not look even remotely as bad as the simulation.
this looks really cool. Could anyone with a phone that has this kinda display say – does the picture on it look good ? Would you say better than on regular lcd ?
Yes, the picture looks a lot better than conventional lcd displays. In fact, after using a PenTile display my eyes often pick out the flaws in conventional lcd screens. The quality of PenTile displays is far superior.
I have a Droid X2 with a Pentile RGBW LCD. I’d say the simulation does a decent job of showing what the cross-hatching artifacts look like, but they are over-exaggerated compared to the real hardware. On my display the cross-hatching is largely unnoticeable(at sane viewing distances) except for green elements.
In general I’d say the display looks very good, but I’d put that down to it being a high-quality panel with high pixel density rather than it being PenTile. Some people are driven mad by the visible artifacts of PenTile, I’m not one of them. You should see one for yourself.
D’oh! I misspelled my website in previous comments. I am not snowblindpeanut. I’ve fixed it.
That colour tint only happens when you don’t account for gamma correction when scaling.
http://www.4p8.com/eric.brasseur/gamma.html
I think it would be fair to say that the SuperAMOLED display in my Sammy Galaxy S2 doesn’t have a pentile layout. I’ve looked at the display very closely, and there’s not the slightest sign of any hatching patterns or weird anomalies of any kind.
From what I heard, the special pattern was used solely to keep the early OLED displays from costing too much.
Now the pentile displays are becoming much rarer (good) as OLED production is ramping up.
Well the new ones are SuperAMOLED+, where the + means its not pentile.
But I have a Nexus One and the screen has never bothered me. I’d gladly buy a pentile device again. Some people just wanna hate. The screen on my N1 is amazing.
The images don’t seem to be gamma-corrected properly. I have a well-calibrated monitor, and the image looks purple even at 100%. But if I apply a gamma correction of 0.55, then the colors are neutral not only at 100%, but at any scale.
Mr. Barrett’s simulator has a major mistake. This simulation is subsampling (turning off ever other green) the wrong subpixel color. It should instead be subsampling the red and the blue subpixels. A further problem exists in that this code does not subpixel render (anti-alias filter) the data onto the subsampled subpixels. The PenTile OLED RGBG processor also examines the red and blue data planes and transfers luminance modulation information onto the fully sampled green subpixels to maintain sharpness of color images. This simulation does none of these things, in addition to having exactly reversed the subsampled color assignments.
Full disclosure: I am the CEO of Nouvoyance, the firm that developed the PenTile RGBG processor technology.