Originally Posted by kohe321
Copying my post from the burn-in thread:
Really cool that someone is doing this test. Will be interesting to see the results as the weeks go by.
My only gripe with the test is that they aren't doing 3 different OLEDs, for example one 2016 model and two 2017 models, so we could see if there were any random discrepancies between them, or if the results were very similar. If I'm not mistaken, permanent IR hasn't ever really been a big worry with IPS or VA LCD panels, and the LCD technology has been used for so many years now that we know roughly what to expect.
There are no differences between different panels and/or different models. At most, there may be a difference between different model years, so having a B/C7 in addition to the B6 would have been useful.
The kind of test they are performing is pretty stressful, so the relative comparison to VA and IPS will be worthwhile.
They are running 20 hours / day 7 days / week or 140 hours per week at 175 cd/m2, which should correspond to something under 20mA/cm^2 (and OLED Light of about 60 initially). Here are my predictions based on the semi-consistent reports of burn-in we've had from several AVSers:
-burn-in of red will become apparent on the 100%/100% yellow bar (Top Left) somewhere between week 3 (420 hours) and week 5 (700 hours). Burn-in might
be visible on a red field after week 2 (280 hours) but if so it will be very minor and invisible on actual content)
-burn-in of red in the 36%/100% 'Video Game' yellow bar (Bottom Left) will take just under 3x the number of weeks to become equally apparent. This will confirm that it is only cumulative hours of qualifying static content (fully-saturated bright yellow/orange/red) that matters and 'rests' or mixing of content type have no effect other than delaying how long it takes to accumulate a sufficient number of cumulative hours to develope visibe burn-in.
-both semi-tranparent logos are going to take much, much longer to burn-in and I won't be surprised to see the test conclude without any signs of burn-in on either semi-transparent logo.
My greatest dissapointment with the test is that they are using these stupid Rtings logos rather than a proper 8-color rainbow sufficiently large to make measurements from (with a meter).
If they had a White, Black, Red, Greed, Blue, Yellow, Magenta, Cyan pattern where each patch had at least a 1" x 1" rectangle sufficient to take a reading from, they could confirm the aging data on a color-by-color and subpixel-by-subpixel basis.
That data would show that 'red within yellow' ages the fastest, all-subpixels-within-black age the slowest (not at all), and 'white-within-white' ages somewhere in between those two extremes (probably about 1/3rd the rate if red-within-yellow, because of the 3x greater efficiency of the (unfiltered) white subpixel).
As it is, they are probably going to use the central area of the screen where random content is being displayed to recalibrate 100% white each week (which will change very little during the test, if at all) and aside from using their eyeballs to check for 'signs of burn-in' on a colored field, they will have no way to measure the % burn-in (reduced brightness) they are seeing...
Hopefully the results will be interesting enough that they decide to repeat the effort more comprehensively in 2018...
Rather than the % brightness versus time assiciated with different levels of current (mA/cm^2), their goal should be to generate curves of % brightness (absolute brightness since they have standardized on 170 cd/m2) versus cumulative time for the various subpixel-colors-within-display-colors (32 measurements total, many/most meaningless, meaning no degradation).
This would confirm that red-within-yellow ages the fastest, followed by red-within-red, and then green-within-yellow (and that blue ages the slowest overall, only white may be slower).
That might push LG to actually deliver static-logo-dimming technology that actually prevents burn-in, rather than just drags out the cunulative display hours needed to develop it