What cause GtG increment when lowering refresh rate ? [GtG Slowdown Effects]

Advanced display talk, display hackers, advanced game programmers, scientists, display researchers, display manufacturers, vision researchers & Advanced Display Articles on Blur Busters. The masters on Blur Busters.
Post Reply
User avatar
AddictFPS
Posts: 212
Joined: 15 Jan 2020, 14:13

What cause GtG increment when lowering refresh rate ? [GtG Slowdown Effects]

Post by AddictFPS » 15 May 2020, 04:05

For instance, AOC AGON AG273QZ in TFTCentral review, 27" TN QHD 240Hz, GtG transition from 150 to 255 take 8.8ms at 240HZ, but 12.5ms at 100Hz, both whit the same overdrive medium, both 100% free overshot.

https://www.tftcentral.co.uk/reviews/ao ... htm#gaming

GtG transition speed depend of voltage regulation applied by overdrive, and if is applied exactly the same at lower frequency, should result in the same speed, but not occurs, and overdrive need to reduce voltage to avoid overshot.

What is the technical explanation of this issue ?

User avatar
Chief Blur Buster
Site Admin
Posts: 8107
Joined: 05 Dec 2013, 15:44
Location: Toronto, Ontario, Canada
Contact:

Re: What cause GtG increment when lowering refresh rate ? [GtG Slowdown Effects]

Post by Chief Blur Buster » 26 May 2020, 21:10

[Moved to Area51]

First, understand that pixel transitions (GtG, aka Grey to Grey) is simply pixels "fading from its old color to the new color", as seen in high speed videos.

Here is a slow-motion video of a common 60Hz display refreshing:
(From the Pixel Response FAQ: GtG versus MPRT)

phpBB [video]


Now you see the pixel transitions in action! Those pixel transitions can be faster or slower depending on a variety of factors such as panel technology, overdrive and display design. Pixels are refreshed one pixel row at a time. Imagine kicking a soccer ball one ball at a time. Different Hz has challenges -- kick more soccer balls too quickly (ending up doing badly) or take time to do running starts to do perfect kicks to each soccer ball. You have a time limit per pixel (soccer ball). Hz changes all of that, so GtG can vary on Hz.

However, we'll answer why GtG speeds up / slows down from refresh rate changes:

GtG pixel transitions in LCD is actual physical momentum

There are multiple causes, but the easiest way to understand GtG is that it's essentially momentum -- basically a pixel coasting from an old color to a new color. (Well 3 color-filtered monochrome subpixels an RGB pixel).

This is true for an LCD (Liquid Crystal Display) -- because those liquid crystal molecules in a pixel are actually moving like a valve to block/unblock polarized light!

Now, we can metaphorically pretend that pixels like soccer balls -- the momentum of kicking a soccer ball from its current position (old color) to the goal line (final color). This makes it easier to explain pixel transitions.

The cause of lower Hz having slower GtG
This is because of the multi-kick effect. GtG over two refresh cycles be like kicking a soccer ball two times. Higher Hz gives the pixel a second kick to get the GtG closer to final color faster.

The cause of overclocked Hz having slower GtG
This is because not enough time for electronics to give a running kick start. Excess Hz means too little time per pixel to give a very strong voltage kick (like having no time to do a running start to kick a pixel faster)

Pixels are like metaphorical GtG soccer balls to be kicked.
GtG is the physical momentum of a rotating molecule (that blocks/unblocks polarized light) in the liquid crystal medium of a LCD (Liquid Crystal Display). As seen in high speed videos of scanout, GtG momentum is fastest just as soon as a pixel row gets a voltage pulse (the pixel refresh voltage). But that's a brief voltage surge, and the pixel continues in its momentum, just like a rolling soccer ball. But that can be re-kicked faster.

Pixels can be "kicked" again
A refresh cycle that comes sooner, means another opportunity to kick the soccer ball faster towards the GOAL line (final color), if the pixel didn't have enough momentum to reach the goal line before the next refresh cycle, the new refresh cycle becomes a brand new soccer ball kick for the pixel.

Goldilocks Hz where GtG is fastest
There's a goldilocks refresh rate where GtG is the fastest. The goldilocks refresh rate is occasionally at max Hz -- but often times, GtG is sometimes fastest at a few percent below max Hz (if overdrive settings is flexible enough), e.g. 144Hz on a 165Hz IPS panel, and 225Hz on a 240Hz IPS panel.

See Soccer Ball Metaphor: Pixels are metaphorical GtG soccer balls to be kicked to understand why pixel response sometimes slows down at too low refresh rates, or too high refresh rates.
Chief Blur Buster wrote:About GtG slowing down when you try to refresh more frequently....
HyperSlayer72 wrote:
27 Mar 2020, 16:20
This might be getting a bit beyond me but it still very much interests me nonetheless. Especially what you said regarding Driver electronics being a limiting factor. Do you have more sources or info to share on what exactly is holding back driver electronics? Ive often wondered what evolution's have taken place that reduce the hardware/processing latency in high hz displays.
Metaphor: Pixels As Soccer Balls
Here are some facts:

1. Not all pixels on a screen can refresh at the same time.
2. Each pixel requires a finite amount of time to "kickstart" the pixel response.
3. Think of each pixel like separate soccer ball or football.

You see screen refreshing in high speed videos at www.blurbusters.com/scanout ... Many screens refresh only a few pixels at a time, left-to-right, top-to-bottom. Like sequencing the days of a calendar, except the calendar week is 1920 days wide (width) and 1080 weeks tall (height), but really fast.

Each pixel has to be kicked one at a time. It takes time for the player (driver electronics) to give that ball a very fast GtG kick, before moving on to the next ball.

Imagie you have 100 balls to kick. Do you kick each one of them as quickly as possible (kick more than 1 ball per second), or do you properly give you a running start to kick the ball faster? You'll spend more time kicking one ball, but you can kick that ball faster (faster GtG)

That's the problem. Speeding up the display driver can slow down pixel response in much the same "Catch-22" way. (Things like resistance, inductance, voltage ramp-up -- are metaphorically running starts to giving one soccer ball a high-velocity kick for faster GtG).

Ball-to-ball time doesn't always correspond with ball speed. However, if you're in a hurry to kick 100 soccer balls in less than one minute, they won't all be powerful kicks (fast GtG). Driver electronics needs to spend more time giving each pixel a HARD kick (Fast GtG).

Now go back and click the links, understanding that pixels are just balls to electronically kick.

Now, imagine that there are over 2 million soccer balls per 1080p refresh cycle, 144 refresh cycles per second. If we dive to semantics, each subpixel are also separate balls -- so over 6 million soccer balls per 1080p refresh cycle! You see, the LCD panel driver electronics are working hard to kick those balls as fast as possible. Fast drivers (less time per pixel) AND fast GtG (fast pixel transition) simultaneously, is very hard to achieve simultaneously. You can do one or the other, it's hard to do both. Just like for real-world soccer balls. If you're in a hurry to kick as many balls as possible, you don't have time to give each ball a running-start kick for high-velocity (fast GtG).

Also, the previous ball (or few) can still be in mid-air by the time you kick the next ball. Likewise, a previous pixel's GtG is still in progress while the display is triggering the next pixel's GtG. Consequently, this means thousands of pixels are in GtG-progress. Like thousands of mid-air balls. To the human eye, this shows up as a "fade wipe" that scrolls from top to the bottom, as a screen "fades" from one refresh to the next -- as seen in high speed videos.

Working against fast GtG is those tiny microwires (the wire grid to each pixel). Tiny wires (for higher-resolutions) means slower-responding electricity pulses than bigger wires (for low-resolutions). That makes it quite a bit harder to kick the balls (pixels) faster. Tricks such as high voltages are used but too high a voltage, and the microwires begin leaking to each other (vertical line / horizontal line streaks found on older LCDs). And so many other tricks in a scaler/TCON, and this is why Overdrive algorithms were invented for LCDs too -- to use a pixel color beyond target, to speed up pixel transitions to target color values. But there are so many issues (ghosting/coronas). Tons of side effects the industry have to whac-a-mole against each other. And it become much harder for some of them when monitors fell below $1000, given the high costs of engineering solutions -- sometimes engineering costs 10x to 100x to improve a panel by just 10% -- the last 10% of improvement is often the incredibly expensive part. At some point, a line has to be drawn and live with the compromises (such as VA dark ghosting, or IPS slower than TN, lower contrast ratios versus fast GtG, etc). That's why highest refresh rates sometimes comes with some compromises, like TN panels and the limited viewing angles...

Now do you understand the Catch-22 conundrum of fast driver + fast GtG?

Fortunately, there's a sort of a display-equivalent of Moore's Law, where refresh rates double approximately every 5-10 years, thanks to all the engineering that goes on behind the scenes.

Welcome to the refresh rate race. ;)
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