I've had some time to mull over the replies I received on my post
wanting to better understand the benefits of watching HDR content with the Custom Curve approach pioneered by Manni01, Arve, and others, vs the "SDR/BT2020 approach using the HDFury Integral. The fact that the available dynamic range is our Projectors is smaller (I used the arbitrary range of 0-100 for the sake of the discussion) than the range of what HDR actually contains (0-200 was my arbitrary figure) is the source of the conundrum.
These are the pertinent and helpful replies (with the bolding mine for emphasis).
Originally Posted by nathan_h
There are a few benefits to sticking with HDR.
One is that you have a higher bit depth and number of steps between black and white. Another is that the grading for HDR content takes the wider dynamic range into account, so you are making better use of the full range of the projector. Yes, SDR might have the same black point and white point, if you want to set things up that way, but it won't have been graded for the larger dynamic range that one gets (even with our limited front projection devices we still have the potential for far more dynamic range than SDR content is graded for) and the nuances possible when processing HDR will be lost in SDR.
Originally Posted by stanger89
The first thing that's very, very important to understand, is that content is either HDR or it's not. Ultra HD Blu-ray (apparently with a few exceptions) is HDR+WCG.
Using an Integral to "disable" HDR, doesn't mean it's not HDR, you can't remove HDR, you can't strip it, what it means is the player has to perform some sort of EOTF manipulation to make it's HDR output compatible with a display using an SDR EOTF.
What I'm trying to say is, it's not a question of HDR vs SDR, it's a question of whether player doing DRC and having the projector use a normal SDR EOTF produces a better result than having the player output native HDR, and having the projector use an HDR EOTF.
To that question, I struggle to see how a player can do a better job* with a "one size fits all", single slider approach vs a completely customizeable EOTF. *I completely understand the possibility that the projector could do a really bad job, but that doesn't seem to be the case.
Originally Posted by stanger89
EOTF is "Electro-Optical Transfer Function", it's the function/curve defines how to map digital codes (electro) to physical brightnesses (optical). A gamma curve is one specific EOTF, the HDR PQ (ST.2084) curve is another.
So, the normal case is the player just sends HDR content (coded with the HDR EOTF, the ST.2084 PQ curve) directly to the display and it deals with it.
If you use something to force Dynamic Range Conversion, then the player maps that HDR content to an SDR EOTF, assumingly a power gamma or BT.1886 curve. This way a display that only understands SDR content will be able to handle it correctly.
In either case, HDR content on the disc needs to be mapped to physical brightness values, it's just a question of how/where that all gets done.
Basically the custom gamma curves are HDR EOTFs that we've created/uploaded to the projector. They're not exactly the ST.2084 EOTF, but they're based on it.
I think asking if HDR or SDR is better for a display under "x" nits is the wrong question, we're watching HDR content either way (UHD Blu-ray), it's a question of who does the mapping better.
And regarding the question about using High Lamp for HDR Content, I asked if this would provide a somewhat wider Dynamic Range, to better accommodate the range available with HDR (e.g. 0-150 using my arbitrary figures):
Originally Posted by stanger89
Oh, absolutely. But lets use real numbers. My SDR config is calibrated for about 14-16fL, or right about 50 nits, I do that in low lamp, aperture = -10. When I made my custom gamma curve (HDR EOTF), I opened everything up and flipped to high lamp, that netted me about 125 nits. So my HDR preset is over twice as bright (peak white) as my SDR preset, that means I've got quite a lot of headroom for highlights, while allowing reference white to be exactly the same.
What I've taken away from these insightful replies:
1. HDR offers a more finely graded range in brightness (more 'steps' or 'levels') vs regular SDR, in addition to the greater absolute dynamic range. This should have some impact on the picture quality, with more 'subtle' gradations within the Projector's available dynamic range. This is basically for HDR mastered content vs SDR mastered content (and setting aside the WCG for the sake of this discussion), regardless of what method we use to view that HDR content.
2. In order to view HDR mastered content on our projectors, with the restricted Dynamic Range, a 'mapping' needs to take place, whether one is using the Integral for SDR/2020, or the Custom Curves recently made available. The question is which does the job better.
3. The Integral approach is less user controllable, as it relies on the strengths/weaknesses of the Bluray player which is doing the initial remapping. Even with the Slider available in some players, this is simply a less precise way of accomplishing the remapping, since the Custom Curves can be customized to each setup, and are dramatically more adjustable or 'tunable.'
4. The use of High Lamp will increase the available Dynamic Range to better close the gap between what HDR offers, and what our Projectors are able to display.
If I've made errors here, or omitted significant issues, please clarify (and I'm ignoring the fact that the Custom Curve approach would also work with streaming content, which is another practical advantage, but not directly relevant to the theoretical questions I posted).
A couple of practical implications:
A. If our Projectors cannot handle the full range of HDR content, then one would do better by using High Lamp to better close the gap. Fan noise aside, should that be problematic in a person's set-up, and setting aside bulb life, wouldn't there be a universal real-world benefit to running in High Lamp for all HDR content (regardless of the method used to map the HDR)?
Edited to Add:
Further comments on this topic indicate that using High Lamp doesn't actually improve the available Dynamic Range, since the Black Floor is raised along with the greater White limit, leaving the same "0-100" range I used in my original post.
But, if one uses a wide open Aperture, along with Dynamic Iris, then the real-world available contrast is enhanced to better accommodate the "0-200" range of HDR. Without DI, no net gain is obtained in Dynamic Range, although it will be brighter overall, which might still be useful.
had suggested that if one's setup is not capable of 100 nits or more, then running SDR/BT2020 with the use of the Integral would be advisable. But if I'm understanding the replies I received and summarized above, then this advice in a sense 'doesn't apply' since it's more a question of which method does the mapping better, since the limited available dynamic range of our Projectors is the same in either case. And the replies make a strong case for the benefits of the Custom Curve vs Integral approach.
If I've made errors in comprehension, or logic, or omitted important variables, please advise. And I hope this is a useful discussion for others (I know it has been for me), as many seem to struggle, as I have, trying to understand how this all works, and in turn, how to best view HDR content on our Projectors.