Chief Blur Buster wrote:It's often surprising even to myself -- how sensitive my readers are! I'm perfectly fine with lots of artifacts but picky with others.MatrixQW wrote:I don't understand Notty. He doesn't accept VG258QR overdrive but accepts XG258Q overdrive.
But it's understandable, and I am capable of understanding -- that's our famously educational Blur Busters popular science to explain to my readers why.
I know people who are particularly sensitive to overdrive differences between monitors. The image, created with help by Jorim, is one of my favourite images for demonstrating ghosting differences of overdrive settings (now in the Pixel Response FAQ). Captured via pursuiting camera to accurately capture motion blur:
Now take a closer look at "Normal" setting which many people claim is "perfect" for them. Look closer. It's still not perfect.
Most are okay with this slight asymmetry. Heck, I'm even okay (as founder of Blur Busters). It's not blatant. But it's not perfect.
Look at the left and right trailing edges -- they're not perfectly symmetrical! Even "Normal" does not have perfect symmetry between the leading/trailing edges. Look at the slight blue tinge at left and the slight green tinge at right edge of the UFO dome. This is a great example of 1ms-human-visibility; since 1ms translates to 1 pixel per 1000 pixels/second. So those discolored pixels from imperfect 1ms, is still human visible. The newer 0.5ms displays helps reduce this even further if you're *that* picky about ghosting.
Now, the problem is that manufacturer claims often appears fluffed -- but they're not fake claims, but are trunctated via an based on a imperfect measurement standard established as a technological compromise. GtG response time is based on an VESA 10%-to-90% cutoff standard, so the ghosting from below-10% and above-90% is still human-visible -- like the faint ghost to the right of the yellow dome.
Nontheless, 1ms imperfections is human visible in this photo! Green tinge versus blue tinge on left/right edge....
TestUFO motionspeed is intentionally standardized at 960 pixel per second for ease of pixel response analysis -- 960 being the closest motionspeed number to 1000 pixels/second that is divisible by common refresh rates 60, 120, 240. It makes it easy to judge the pixel response of a pursuit camera picture -- even for 1ms "Normal", you can see about maybe 4 or 5 pixelwidths of ghosting to the left of the yellow dome, most of it below the 10% threshold. You see about 1-2 obvious pixelwidths of discolored blurring (ghosting) and 3-to-4 faint pixelwidths of disclored blurring (ghosting). That 1ms GtG to 90% is still like at least 5ms GtG to 100%.
Nontheless, that out of the way, the bottom line is that 0.5ms GtG (not 0.5ms MPRT) claimed monitors will generally have even fainter ghosting than this. Some people are extremely sensitive to this.
Oh, and it gets more complex with variable refresh rate that does not have good overdrive. Sometimes you get bad ghosting only at certain frame rate ranges; it might be like "Extreme" at some frame rates and fading closer to "Normal" when framerates increase. The high priced NVIDIA G-SYNC module is designed exactly to try to avoid ghosting asymmetry during varying frame rates; as part of the price of the G-SYNC premium. That said, some FreeSync monitors are good, and pixel response sufficiently fast enough that it doesn't need overdrive help to remotely reach at least "Normal" quality. Disabling overdrive completely is often not an option, because even "Normal" (light overdrive) for many is so vastly superior to having no overdrive at all. But not everyone like even the slight discolorations at the arrows... ouch.
Also, the curve shapes are not the same. You can have very good 10%-to-90% GtG that completes the last 10% quickly or the last 10% slowly. Basically 1ms(10->90%) with 5ms(100%) or 50ms(100%) -- the shape of the curve matters a lot.
Even what happens below 10% or above 90% (still human visible) might be slower or quicker in different models of exactly the same GtG 10%->90% -- a single number doesn't tell you the shape of the curve. And the curveshapes are sometimes totally different for totally different pixel-color pairs -- VA pixels have very interesting curve shape differences for dark colors than for bright colors. Ouch. No wonder a single number can be blamed as manufacturer fluff when it's merely trying to be an average that misses a lot of nuances.
It's not surprising that many people claim the manufacturer is lying about response time (that's often a bit over the line.... I'd say "I wish the industry would come up with better measuring standards instead"...). Pixel response is sometimes a big unmeasured problem, that is often left to others to measure, and no single site can ever go the comprehensive depth it so deserves -- since it ends up becoming a lot of "TL;DR" stuff of complexity. So the industry likes to use simpler single numbers when marketing.
Your vision is not the same as the next person. You might need eye glasses. You might not. You might see better in the dark than the other person. Or you might be color blind. (Or simply less sensitive to colors). The bottom line is that everyone sees slightly differently than the next individual. This goes for many kinds of display artifacts -- some see them blatantly well, others don't.
The ghosting may apparently be 10x more visible to the next person than the other. Even how it looks in a pursuit camera photograph -- may stand out more to another person than others. I've politely pointed arrows at the asymmetry -- it's subtle to some eyes but more visible to other eyes.
This relates to other kinds of sensitivities (temporal, spatial, colorspace, etc). Just like some are sensitive to tearing. Or sensitive to stutters. Or sensitive to bad viewing angles. Or sensitive to motion blur. Or sensitive to lag. Different people have different priorities in a monitor.
TL;DR: Pixel response, ghosting, coronas, GtG, MPRT, etc -- creates a huge Pandora Box of display topics that invites plenty of vigorous online debate, sometimes as vigorous as the historical "Human can't see 30fps vs 60fps" debates of yesteryear.
This is just masterclass. Thank you so much for this again Chief. Your knowledge is something astonishing and Im so happy you created this website and teach us everyday. Simply amazing!