If I choose HDR, Reference, BT2020, or P3 (from memory) it engages the filter. I can't choose whether to use the filter or not, it's connected to the colour mode. I think the 9900 is something like 93% P3 coverage. I know I can't hit 2020, but it'll do what it can.

Then a second profile for HDR with the iris wide open and a P3/2020 target space. This profile will also have a slightly closed iris, to taste I guess? The SDR one should have a limited gamut but HCFR should tell me that the colour range is accurate to 709. My goal is then to come out of this with two profiles in the PJ, one for SDR and one for HDR. It's not too bad as I can sit in another room and run the cal from there so I don't need to spend hours in a batcave watching test patterns. I'll get the network cable out one night this week and have a play. Ideally we would want a 3dlut for 2020 and everything fits inside that but it's difficult to do a 3d lut when the display doesn't get close to covering the colour space, I don't think displaycal can do that.Ī fundamental gap in my knowledge, it seems. Probably also there will only be smpte 2084, no gamma and everything will map into that.īut madVR does a really good job of mapping one into the other, including into a 3dlut. One day in the dim distant future there will only be 2020 and all (pretty much) visible colours can be represented within that, so every video player will just map whatever olde worlde video into that and we will not have to care. Almost no colour if you send 2020 and the projector thinks 709. You have to match the projector colour mode (colour space) to the colour space MadVR is using, 'this display is calibrated to' otherwise toytown colours if MadVR sends 709 and the projector thinks 2020. The number is the same but the meaning is different for the different source gamut and what gets displayed is different depending on the projector mode, all for the same RGB numbers. In a 2020 colour profile the colours are far more saturated and the projector can't get anywhere near. If the projector is in User1 set to 709 (same as natural I think) that means 100% 709 red, green or blue, in Reference (P3 I think) that is a stronger red or green, blue is less different. The input signal is the same 0-255 (1024 for 10bit), it knows nothing about gamuts it's knows only RGB. No, they interpret the incoming signal differently. Thankfully there's nothing to do for HDR, any user curves are derived from the base 2.2 gamma curve. I just do high and low lamp wide filter, that takes long enough. I was breaking my 709 colour cal by trying to do a 2020 one on top, doh! So colour wise you need to do filter and no filter, lamp high and low and the iris steps for each one but it's ok to only do 709 and P will fall into line. Saves some time and explains why my 2020F looks better since I started using 709F instead of 709NF, I never was able to get 2020F autocal to work usefully, Dominic tells me I can't, it's always bad because autocal only cals 100% primaries (which the JVC can't do in 2020) and I don't need to anyway! So if I colour autocal 709F (wide filter) it also corrects P3F and 2020F as they are all derived from each other. What I recently learned is that all the colour profiles are derived from each other in a similar way.

I only use filter on wide and lamp high lamp low so I only need two gamma autocals, both done at iris 0 smallest zoom. So you probably know about the gamma autocals, you need one for every combination of lamp power, P3 filter, cmd. I recently had some back and forth with Dominic Chan, he told me some useful things about how the colour profiles interact that saves a lot of time.

Displaycal video mode