lexlazootin wrote:My god, i just tried to do the same test on my screen and you're right, it sucks. I think someone needs to figure out a new way of doing it because i'm flinging my head with the alien just to keep track
240hz white background - 9 ppf - 5.1ms - 196 mmcr
240hz black background - 15 ppf - 3.1ms - 327 mmcr
average - 4.1ms - 261.5 mmcr
Mangemongen it should be 1 frame like every other monitor.
Spot on!
There are differences in response-time-acceleration going from black-to-white, and white-to-black. Some LCD's go from black-to-white faster than white-to-black. Therefore, a more accurate MMCR can be obtained by averaging the two values.
So if you want an accurate MMCR value, measure both values, then average the two values.
When the old MPRT test pattern was created, it was designed in the era of 60Hz and 120Hz LCDs -- the 165Hz-and-faster LCDs had not been tested yet at the time. Also, back at the time, response time acceleration differences interfering with MPRT testing were not widely known. It's extremely difficult to design an MPRT-measuring test that relies 100% only on human vision. You're trying to actually measure milliseconds by human vision alone, via the motion blur trail length effect.
That in itself, that it is possible at all, to create a test pattern that allows the
unaided human eye to measure 4.1 milliseconds
by human vision alone -- is still impressive. Basically you're actually, measuring, by TestUFO pattern, the size of 4.1 milliseconds worth of motion blur, by human eye alone, to an accuracy of +/- 0.1 milliseconds,
in a web-based motion test pattern! Ponder that for a moment...
That way, it won't be a comfortable or easy test pattern, but the fact that it's possible to measure MPRT by unaided human eye to +/- 0.1ms accuracy -- on many screens -- (if you average the white and black values) -- is quite a feat -- and impressive, even to myself.
That said, I have subsequently discovered new ways to slow down the motion tests for 240Hz monitors. I'll attempt to upload a new MPRT test sometime later this winter --
once the current Blur Busters upgrades are completed.
We've later found out that to get an accurate MPRT measurement, you must measure both the white & black values, and average the two together, as above. Done this way, you get ~0.1ms accuracy as long as there's no strange overdrive sheninigians going on.
On about 80% of LCD's I've been able to get +/- 0.1ms MPRT measurement accuracy via this
http://www.TestUFO.com/mprt test.
The other 20% of LCD's behave so strangely, or doesn't have proper overdrive, that it's not easy especially if the black-to-white blur trail looks very different from the white-to-black blur trail -- which then means there's a "subjective judgement" error factor trying to align the checkerboard appearance when one edge of all squares in the scrolling checkerboard is much blurrier than the opposite edge. This hugely asymmetric GtG effect can add a +/- 1.0ms error margin in my experience. You definitely need well-balanced response-time-acceleration that's accurate for all GtG combinations, where possible, and ideally you want to try to max-out your contrast a little, so black is full-black, and white is full-white (So you don't want to digitally compress your greyscale range during the MPRT test) -- for proper MPRT test pattern testing.
The better the overdrive, the more accurate the MPRT test becomes, since you need actual GtG LCD response time (to ~90% completion of transition) to be properly far less than a refresh cycle, in order to have the MPRT test properly work.
You can make MPRT easier (slower scrolling) by using smaller size squares in the scrolling checkerboard, but the MPRT measurement error margin definitely goes up significantly.
MMCR is a number roughly equivalent to Samsung Clear Motion Rate (CMR) and LG Motion Clarity Index, etc. These numbers are simply inverses of the scientific MPRT measurement. e.g. 240Hz = 1/240sec MPRT = 240 MMCR = 240 Samsung Clear Motion Ratio = 240 LG Motion Clarity Index. If you are wondering why the heck the TV manufacturers advertise those "fake refresh rate numbers", that's simply an inverse of MPRT measurement, and an approximation of expected motion blur for framerate-equalling-refreshate situations (usually, alas, "interpolated" in those televisions). But what we have here, the XL2540, is a bona-fide true 240Hz native refresh rate display!
Yes, the
www.TestUFO.com/mprt pattern is hard and fiddly. While a technical breakthrough in enabling unaided-human-eye MPRT measurements to sub-millisecond levels, it is crap for ease-of-use. There will be a much easier MPRT measuring test pattern in 2017.