Originally Posted by stanger89
I absolutely will make my own curve, I just haven't had the free time to do it yet. And also I'm not asking you to make a curve for me, I am simply discussing the theory of making such curves. Specifically what the target of that curve should be.
Taking a step back from the practical limitations that may exist in the JVC projectors. Do you disagree that an "ideal" mapping of HDR for a projection setup (which can nit 100nits+) should have 100 nit content displayed at 50 nits or so? And if so, what should it be and why?
I entirely agree with the theory, and this is definitely what I would do if I had access to Dolby Vision Cinema content, but I disagree in practice for all the reasons already mentioned because what we have is content mastered to 1000nits and up, whether we like it or not. If I display this content in HDR, I want to keep the benefits of HDR (wider gamut, less banding as well as high dynamic range). I don't want to force a square peg into a round hole and lose all the benefits.
Because from my experimentation, the smaller the multiplier (the lower the target), the brighter the curve (as expected) BUT the less saturated the picture is. So yes, you're going to be able to play HDR content in a way that will look bright enough, but you are losing ALL the benefit of the HDR master. What's the point?
I just got fed up so I made a quick curve (it still took me an hour of a time I don't have more than you do!). It does exactly what you want. I'm sure all the brightness deprived users will be very happy with it. Problem: it's very undersaturated, as expected, and all the highlights are compressed into a minuscule bandwidth. You could of course start the roll off earlier (I started it after 60%, so after 240nits). In conclusion, yes, you can play HDR that way on a brightness deprived screen, but no, in my opinion, it's not better than SDR BT2020, at least with the tools we have at the moment. I hope Calman will come up with an HDR10 Projector calibration workflow which will allow us to do better than what we can achieve simply with the multiplier. It's better than nothing, but it's not ideal.
I make the curve available for people to check it, but please check it against SDR BT2020 before thinking it's great. Sure, you can pump up the color control like mad and whatever, but it's not the way to do it with the tools and the consumer content we have.
If people, through experimentation, come up with tips and advice to improve this, I'll be all ears. But in the meantime, please stop from making theoretical suggestions until you've put the hours in and checked for yourself. There is no free lunch and there are pluses and minuses for each option. Brightness is not the only factor in HDR. You also want a wider gamut, no?
My recommendation is still not to target lower than 600-800nits. Below that, I would go for SDR BT2020. I personally like the brightness when targeting 1000nits on my screen, but I get more saturation targeting 1100nits, which is still bright enough for me. Hence this is my preferred target for now.
Anyway, 200-400-4000 has been added in the first post of the JVC Autocal in the high lamp section. It doesn't look as bad as I was expecting it to (saying it would be black and white was an exaggeration
), but I certainly won't be using it.
Originally Posted by stevenjw
Just do the rest of us a big favor and NOT leave this thread. Use the ignore function or whatever it takes for you to stick around.
No worries, I had not planned to leave the thread, I just got tired of the theoretical debates, from people who know better without putting the hours in the experimentation of this new tool.