MatrixQW wrote:To get to the lowest input lag and response times means we must use the highest overdrive setting, wich noone ever uses, right?
Not necessarily. The 0.5ms may have been measured at normal overdrive setting.
Very good fast native GtG can produce better-looking overdrive with less overshoot artifacts.
The existence of native GtG that is extremely fast, makes it easier to add overdrive without adding visible ghosting/coronas.
There are still human-visible improvements to clean 0.5ms GtG versus clean 1.0ms GtG when it comes to ultra-high refresh rates.
GtG doesn't matter at 60Hz when 1ms is only a tiny fraction of 60Hz (1/60sec = 16.7ms), but it becomes a massive chasm at 480Hz (1ms is 50% of a 480Hz refresh cycle). The higher the refresh rate, the more visible GtG problems becomes.
240Hz is the refresh rate where going faster than 1ms becomes useful, and for future 480Hz it becomes essentially much more necessary.
Blur Busters Law
1ms = 1 pixel of motion blur per 1000 pixels/sec.
Remember, at the standard TestUFO speeds (nearly 1000 pix/sec), 1 pixel = 1ms
Observe the symmetry between left and right edges.
1. In OFF, there's ghosting at the left edge. (Roughly 8ms worth = 8 pixels of discolored blur as "ghost")
2. In Extreme, there's a bright corona at the left edge (roughly 8ms worth = 8 pixels of discolord blur as "corona")
3. In Normal, it's almost symmetric. But not perfectly symmetric because 1ms GtG isn't perfect enough to make it perfectly symmetric
Pixel response limitations often last for approximately 2 refresh cycles on TN
(can be way more on some panels, especially for dark colors on for VA panels).
At 1920 pixels/sec, the 1ms GtG limitations start to become much more human-visible, and this is where 0.5ms GtG can help make the left/right edges of fast horizontal motion much more symmetric looking, but even 0.5ms still won't be the final frontier (going better than 0.5ms will still be needed in the continued Refresh Rate Race 240Hz->480Hz->960Hz of the 2020s/2030s/2040s).
--> That's why I school those "0.5ms vs 1.0ms" naysayers. <--
<Blur Busters Technology Pandora Box="OPEN">
Complication: Variable refresh rates. Also, GtG often remains constant during variable refresh rate operation, so ghosting/corona discolorations can vary during framerate changes. Dynamic overdrive (variable overdrive) on variable refresh rate displays need to try to maintain the blur symmetry at ALL refresh rates, which is extremely challenging, especially considering it has to predict overdrive for the NEXT frametime, not the previous frametime. Poorly-tuned variable refresh rate monitors have extremely bad variable ghosting artifacts during varying framerate situations. Good ones (especially ones calibrated by NVIDIA for G-SYNC) often have superior consistency in overdrive at all framerates, but even 1ms GtG starts showing major limitations. Turning off overdrive (3ms GtG) is dead-on-arrival, as that sometimes creates 6 pixels of asymmetry at 2000 pixels/sec, so overdrive needs to be enabled to equalize the blur and make it closer to a symmetric blur rather than an asymmetric blur. Alas, if only we had 0.1ms GtG pixel response WITHOUT overdrive.... (OLED, hint, hint).
</Blur Busters Technology Pandora Box>