Pixel behaviour in sample and hold, in unchanging image.

Advanced display talk, display hackers, advanced game programmers, scientists, display researchers, display manufacturers, vision researchers. The masters on Blur Busters.
User avatar
Chief Blur Buster
Site Admin
Posts: 6512
Joined: 05 Dec 2013, 15:44

Re: Pixel behaviour in sample and hold, in unchanging image.

Post by Chief Blur Buster » 15 Dec 2014, 16:27

blargg wrote:Doesn't the LCD hold the voltage between passes?
Correct (active matrix transitors keeping voltage constant to pixel).. But the LCD pixels can stick while it's momentuming to its final color value. They are prone to retention, and they tend to resist, get stuck (not sure of physical phenomenae; probably stiction in the molecules? not sure). That's why inversion exists. The voltage alternates to positive, negative (LCD inversion explanation), this kicks the LCD pixels out of its tendency to image-retention, pushing it closer to the final color. I believe, for the most part, that's why subsequent refresh passes "kicks" an LCD pixel ever closer and closer to its final color value. There may be other reasons, but this is the main explanation I can think of. There could also be decay in the active matrix transitors that needs to be re-refreshed (like a DRAM refresh pass), as TN LCDs will gradually fade to white when there's no refreshing -- no repeated voltage-resetting passes at the active matrix transitor gates...

A little known factoid: This is why I had to make modifications to http://www.testufo.com/flicker to prevent image retention (aka LCD burn in) of a forgotten browser window -- the flicker test causes image retention after a minute or so. Because flicker patterns often 'defect' LCD inversion ability to erase image retention. Every 10-20 seconds, it adds a repeat refresh (e.g. black-white-black-white-black-black-white-black) to reverse the phase. Even so, if you drag your browser window half an inch up/down once every 5 second at http://www.testufo.com/flicker on certain monitors there will be a faint evidence of the old boundaries.

I should create a TestUFO LCD test pattern that demonstrates image retention (inversion-defeating animation, such as text flickering for 30 seconds every other refresh cycle), followed suddenly by a solid image (to show LCD image retention) followed by an image-retention-erasing flicker sequence (black-black-white-white-black-black-white-white-black-black). Flickering items every refresh tend to defect current LCD inversion algorithms, while flickering items every other refresh tend to undo the image retention quite quickly.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: Pixel behaviour in sample and hold, in unchanging image

Post by spacediver » 15 Dec 2014, 16:59

Thanks for the very detailed reply. I've split the topic as you suggested.

Ok, let me sharpen my question, to see whether I'm understanding correctly.

Suppose you have a sample and hold, active matrix display, and the display consists of only one pixel. For simplicity, assume no subpixels or color filters. Also assume no overdrive.

Now say you instruct the video card to render the following pixel values (where 0 is min luminance and 255 is peak white)

Frame 1: 0
Frame 2: 255
Frame 3: 255
Frame 4: 255
Frame 5: 255
Frame 6: 255
Frame 7: 255
Frame 8: 255
Frame 9: 255
Frame 10: 255


My question is this: Are there voltage kicks between frames 2 and 3, and between frames 3 and 4, and so on?

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: Pixel behaviour in sample and hold, in unchanging image

Post by spacediver » 15 Dec 2014, 20:13

In liquid crystal pixel cells, it is only the magnitude of the applied voltage which determines the light transmission (the transmission vs. voltage function is symmetrical about 0V). To prevent polarisation (and rapid permanent damage) of the liquid crystal material, the polarity of the cell voltage is reversed on alternate video frames.
Just so I understand this correctly:

Say that 0 V represents a molecular orientation associated with maximum blocking of light (video level 0). If 0.5 V means that molecule rotates by 90 degrees in one direction and allows maximum transmission of light (255), does that mean that -0.5 V means that molecule rotates by 90 degrees in the other direction (relative to 0V), and also allows maximum transmission of light (255)?

If so, then I think this helps me with my original question. It means that these polarity shifts are what cause voltage kicks between frames even when the image is the same from frame to frame.

User avatar
Chief Blur Buster
Site Admin
Posts: 6512
Joined: 05 Dec 2013, 15:44

Re: Pixel behaviour in sample and hold, in unchanging image

Post by Chief Blur Buster » 15 Dec 2014, 21:48

spacediver wrote:My question is this: Are there voltage kicks between frames 2 and 3, and between frames 3 and 4, and so on?
All pixels are refreshed every refresh cycle, even if the color values haven't changed. It has to, or the pixel would decay or stick (without voltage inversion). The stepped behavior of GtG over multiple refresh cycles, also can be more visible on panels that decay fast on loss of power (ones that would fade to white within one second) and fainter on panels that decay slowly on loss of power.

If you mean voltage kick as overdrive voltage, then no -- that's typically only done when the pixel color value changes, so only between frames 1 and 2 in this example. But each pixel (aka the gate of each active matrix transitor behind each pixel) is pulsed every single refresh cycle, no matter what. Otherwise, LCD pixels will stick or decay (fade to white). A more accurate terminology might be a "push the pixel color value ever closer and closer to their final color value".
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

flood
Posts: 897
Joined: 21 Dec 2013, 01:25

Re: Pixel behaviour in sample and hold, in unchanging image

Post by flood » 15 Dec 2014, 22:20

so basically, lcd's suck and are quite complicated...

User avatar
Chief Blur Buster
Site Admin
Posts: 6512
Joined: 05 Dec 2013, 15:44

Re: Pixel behaviour in sample and hold, in unchanging image

Post by Chief Blur Buster » 15 Dec 2014, 22:41

flood wrote:so basically, lcd's suck and are quite complicated...
Surprisingly so.

All displays have their engineering quirks....even OLED.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: Pixel behaviour in sample and hold, in unchanging image

Post by spacediver » 15 Dec 2014, 22:52

Chief Blur Buster wrote: All pixels are refreshed every refresh cycle, even if the color values haven't changed. It has to, or the pixel would decay or stick (without voltage inversion).
I see, that actually is consistent with the image from this document:

http://www.vpixx.com/datasheets/Pixel_r ... mation.pdf

If you look at the third figure, you can see that there's a rise and fall time for the pixel that can't be explained by the scanning backlight alone. If the molecules in those pixels were fully oriented for maximum transmission, for entire duration of those frames then the rise and fall times would be virtually instantaneous as the backlight scans on and off. But you can see a clear stepping function. I'm assuming this stepping function (which is symmetric on both the rise and fall) represents a double kick of voltage?

Also, I can confirm that when I tested the VPIXX display with my own motion blur experiment, there was a faint and discrete ghost, which seems to correspond to that stepping function.

User avatar
Chief Blur Buster
Site Admin
Posts: 6512
Joined: 05 Dec 2013, 15:44

Re: Pixel behaviour in sample and hold, in unchanging image

Post by Chief Blur Buster » 15 Dec 2014, 23:00

spacediver wrote:I see, that actually is consistent with the image from this document:
http://www.vpixx.com/datasheets/Pixel_r ... mation.pdf
Actually, over those timescales (pulses ~8ms apart), I don't think that's the GtG stepping (unless the Viewpixx pulses the panel severak times per refresh cycle, which I don't think it does). That would be a far smaller error margin than that. What I see is probably backlight diffusion behavior of a progressive-sequential scanning backlight. From the looks of the graph, it looks like there could be about 16 rows of LEDs in the Viewpixx, and each of them are progressively illuminated, and the stepped effect is light leakage of backlight diffusion from progressively-further-away scanning backlight segment illumination.

P.S. spacediver -- sent you PM
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: Pixel behaviour in sample and hold, in unchanging image

Post by spacediver » 15 Dec 2014, 23:11

I see, that makes sense.

As for the voltage inversion, does this actually flip the molecules? There's something that's bugging me.

If voltage inversion is done to protect the molecules from sticking, then I'm assuming that each time the voltage polarity flips, then the orientation of the molecule also flips. In other words, it might flip from +45 degrees to -45 degrees (and I'm assuming that +45 and -45 orientations represent equal amount of light transmission).

But if this is the case, then why would each successive flip bring the pixel closer and closer to its target? Wouldn't each flip mean that the whole thing starts from scratch each time it flips?

If this question isn't clear enough, let me know, and I'll rephrase.

User avatar
Chief Blur Buster
Site Admin
Posts: 6512
Joined: 05 Dec 2013, 15:44

Re: Pixel behaviour in sample and hold, in unchanging image

Post by Chief Blur Buster » 16 Dec 2014, 00:27

I'm curious for more information too, since I'm not familiar with how inversion affects orientation of liquid crystal molecules. But I don't think their orientation changes, except due to the imprecision of the positive versus negative voltage (as explained in Techmind, it's hard to keep the voltages exact...)

Try googling on Google Scholar and other sites, or looking through patents (e.g. I found this gem http://www.google.com/patents/US7893910 ...)
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

       To support Blur Busters:
       • Official List of Best Gaming Monitors
       • List of G-SYNC Monitors
       • List of FreeSync Monitors
       • List of Ultrawide Monitors

Post Reply