Originally Posted by losservatore
This is the pixel matrix of the Vizio P series ,I'm not sure if this reveal if the panel is 10bit or 8bit + HI FRC,but I really don't care as long as it have a great PQ.
Quote from Vizio CTO Matt
McRae on the Vizio P thread
(from Matt regarding P series panels and bit depth:
As you can imagine... we are now getting close to secret sauce :-) I can tell you that the panels are true 10bit (not 8bit + FRC dithering etc). And I can tell you that to reach this color fidelity we modified LED phosphors and chemistry AND spec'd custom color filters.)
enjoy watching the pixel structure.
I'm trying to understand a bit better how both LCDs and OLED panels are considered native gamma 2.2 instead of linear, and so far, I have two reasons for thinking so:
One, the subpixel structure itself, if there were enough domains and they were the right relative sizes to correlate with the power curve, they could simply turn each one fully on or off and be done with it. But, even for true 10-bit panels here, there seem to be at most eight subdomains, of different sizes, but that doesn't make the displays native gamma, but something else which is technically non-linear but not at all gamma.
What I'm saying is, it must then be the case that each subpixel domain, or combinations of domains, can be attenuated in an analog way, and that the final number of bits and the encoding of those bits is relative to the number of lanes, 3, 4 or 5 to 18-bit, 24-bit, or 30-bit capability, plus the response curve of either the DACs or the eletrical response to voltages, in which case the DACs are likely linear DACs (as opposed to logarithmic), and their desirable non-linearity is due to the way the liquid crystals respond to voltages.
Has anyone figured out, or leaked the actual formula for how one takes, say, a give 10-bit PQ-encoded value, and show what values are sent to the different sub-pixel domains, and then how those domains in turn control the final transmissivity of their subpixels?
I'm also curious how OLED subpixels work, is there a similar way to determine the native bit depth, like are the sub-domains there too, or are they simply controlled individually via a non-linear DAC or inherent voltage response. I highly doubt there are any real 1:1 DACs that can natively deliver 10-bits of PQ encoded data, and I'm fairly sure I've read virtually all OLEDs are gamma 2.2 native, not PQ native, so the processing involved could in theory result in visible banding. Unless the Gamma 2.2 subpixels were encoded to, say, 12-bit or 14-bit gamma 2.2, then I could see 10-bit or 12-bit PQ being shown faithfully.
It does seem to me that having differently sized sub-domains in each LCD subpixels could easily be used to increase effective bit depth, roughly following a power law due to them being different sizes, so that's pretty smart. I also wonder if FRC is typically applied to only the smaller sub-domains since they might transition faster than the larger ones, due to having less liquid crystals to rotate inside a smaller volume. So even if the smaller four sub-domains are tied together electrically, splitting them up into four identically controlled sections could maybe boost their overall performance and G2G response time. I wonder also if the G2G response time is directly related to the smaller size of the smallest domains. We know LCD panels can be overdriven but then the crystals overshoot the mark and result in some haloing or ghosting.