Overshoot from 0% seems to be intrinsic to WOLED and Rtings.com has been posting measurements since the first WOLED generation which clearly shows the overshoot: https://www.rtings.com/tv/reviews/lg/cx-oled
Looking at the 0-20% measurement, the overshoot looks to be about 2.85% luminance (~20% / 7), while looking at the 0-100% measurement, the overshoot looks to be about 2.5% luminance (~1/5th x 100%/8), so the overshoot appears to be a fixed luminance % in the best case and increasing in absolute value at lower luminance steps out of 0% in the worst case.
A 2.85% overshoot to 20% luminance is probably not going to be visible, since it’s less than 15% of the luminance increase / step, but assuming that same overshoot of 2.85% remains at a luminance step from 0% to 10% or even 5%, we’re talking about an overshoot of 28.5% to 57% of target luminance, so a luminance overshoot flash lasting 10ms in a dark scene will be visible, at least if several pixels in the same area overshoot at the same time.
From what I can tell, aside from ongoing panel-lottery issues associated with uniformity and near-black DSE, this luminance overshoot issue is WOLED’s biggest Picture Quality bugaboo.
Hopefully with LG now focusing more attention on the issue, they’ll find an acceptable workaround.
One idea that would probably work would be to break up the correlation - since they have some latitude on what luminance level to activate the white subpixel (which seems to be the primary culprit), breaking a thing up so every 10th pixel (or whatever) turns on it’s white subpixel at a different time / luminance threshold would go a long way to avoiding entire flashing / overshooting areas / lines / contours which are large enough to draw attention.
I hope LG finds an acceptable fix for this issue since a dark junkie like me considers any black-level anomalies a showstopper when selecting a new TV...