One additional bit of information that I had forgotten about: from what I've seen, some of them were only visible from certain angles, and disappeared otherwise. I'd venture a guess that supports the 'micro-sized dust or dirt in the liquid crystal'.Originally posted by: xtknight
I don't even know how that's possible. It could be micro-sized dust or I suppose some dirt in the liquid crystals. Or, it could be the fact that each subpixel is divided into anywhere from 2 domains (DD-IPS) to 8 domains (P-MVA/AMVA) (or not divided at all). I had thought that S-IPS screens were only single domain, though. My theory is that somehow one of the cells in those domains is malfunctioning (stuck liquid crystal or poor LC solution, likely possible). Actually, S-IPS screens require two transistors at the backend to drive them (which is why they suck more power), so that could be it also. It may have nothing to do with the cell domain configuration.
It all seemed very strange to me as well, that's why I brought it up, but there seem to be comments on the Net implying such defects are at least as common as fully-stuck subpixels. It could be dependent on technology, as you said: on multi-domain panels, it does indeed make sense for just a subset of the cells in a subpixel to be stuck; on such configurations, it even seems less likely for all cells to be stuck than only one or a few.
So, in the end, there may be different reasons for them depending on panel technology. This is all speculation, of course, but I think these thoughts are worth passing back and forth.