Originally Posted by lovingdvd
First, you should know that any post I am responding to about this complex HDR curve discussion, I have read at least a dozen times before responding.
I read it again just now a few more times per your recommendation to revisit it. Unless I'm missing something, it doesn't address my question. From your message I gather that we have an option (?) of setting black by using 64 as a reference (and targeting, 68 blinking) or targetting 77, 81 blinking). But my question is - which is the right one to target? When would you target one over the other? Now I may have missed that from another post at some point, but unless I am really overlooking it, its not in the post you referred me back to. Thanks!
I know you have read the post, and probably re-read it, but I'm asking again: please re-read the post.
In that post, I am giving two options for setting the black point. I am explaining what each one does, for which titles. I'm asking people to try each one, with different types of content, and see if they can see the improvements I see for one, and no downside. I ask if they can see any improvement with the other option, which I don't. I'm not going to repeat this in detail everytime one doesn't understand, it would take me ages! However, because you're a good friend and because you're not the only one, I'm going to expand a bit, but you still need to re-read the post in question (it called setting brightness in HDR and it's linked to in the first post of the Autocal thread in the HDR gamma section if you're not sure which one it is)...
[EDIT: Here's the link: https://www.avsforum.com/forum/24-dig...#post51346873]
and a bit more info here: https://www.avsforum.com/forum/24-dig...#post51359041]
Until I get enough detailed feedback about this, from users following specifically what I've asked about in that post and doing exactly the tests I've asked, there is no answer to this question.
There is a "theoretically" correct way to set the black point, and that's resolving down to level 68 (black at 64). This raises the black point in most titles (those mastered to 0.005nits, especially those mastered to 1000nits, such as Lucy). It should increase shadow details for clips mastered to 0nits, such as Hitman 47 or the Revenant, but I as far as I can see, it mainly raises the black floor in 0.005nits titles (for example, it makes the black bars visible in Lucy), but it doesn't add any shadow detail that I can see in 0nits mastered titles.
What I'm suggesting to do is to use a theoretically incorrect way to set the black point, and to resolve only down to 0.005nits, so level 81 blinking, black at 77. This means that for all titles mastered to 0.005nits (about 75% of all the titles to date, including all titles mastered to 4000nits) we only resolve down to the black levels of the grading monitor. By not trying to resolve lower, where there is no content and no further shadow detail, we don't raise the black floor unnecessarily. This is especially visible in a title like Lucy. Take any of the fade to black at the beginning, you should see the black floor going down significantly. There shouldn't be any downside doing this for all the titles mastered to 0.005nits black. HOWEVER, in theory, when we do this we crush shadow detail for the titles mastered to 0nits, as for these titles there is shadow detail below 0.005nits. But when I switch between the two curves for these titles, I can't see any black crush or visible loss of shadow details. I've used Hitman and The Revenant, both mastered to 0nits black, to try to see if using the "correct" curve would increase shadow detail, but to my eyes it doesn't.
This is why I'm asking other to try these two options after having set black brightness correctly for their setup, and report if they see the same advantage with option 2 with 0.005nits titles, and no drawbacks (i.e. option 1 doesn't bring back any visible shadow details on titles mastered to 0nits).
: when switching between the two gamma curves to compare black floor / shadow detail, don't use the menu, as this makes it much more difficult to see the difference (plus it opens the iris wide for those that use the DI). Cycle between the gamma curves using the gamma button on the remote. This only displays the active gamma in the upper right corner and is less damaging to the black levels as your pupil can remain closed, so you see more shadow detail.
Until I get this feedback, with detailed timecodes as requested where a 0nits curve seems to improve shadow detail, I'm not going to make any recommendation. I'm using option 2 in my Unicurve because I can't see any downside to resolving only down to level 81, but that's in my set-up. In others, maybe there is a downside, which means some would have to use two curves, or only the correct one and live with the raised black floor in some titles.
By the way, I checked the curve and the difference between a 68 and an 81 curve is just one click at the 5% control point in the lowres gamma curves I exported. I'm sure more points are affected in the original curve, but still, it's not a lot, which explains that while it might improve the black floor in many titles, it doesn't lead to a significant loss of shadow detail in the others.
Originally Posted by Dave Vaughn
Here's what I came up with tonight:
Max bright: 800nits
Ref white: 100nits
Black brightness: 0.005 (resolving down to 0.005nits/level 81 on my unit for black at level 77)
Soft clip start: auto (335 effective if you're not using wip yet))
Hard clip: 4000nits
Soft curve type: 0
Soft curve gamma: 1
If I set to .01, I can resolve down to 68 on my unit, making black level at 64. Not sure if I should stick with 77 or go down to 64 though. What do you think? I tested my black levels and I haven't raised the black floor at all.
Please read the above as well as the post mentioned, do the requested testing and report back
Originally Posted by krichter1
And so I'll ask again... can we only import your 11 point gammas or can any of the "universal" ones you created (even 256) be used Manni?
If you don't use Arne's tool, you're stuck with the low-res version at 11point. If you use Arve's tool, you get 256points, which might or might not make a visible difference, depending on the curve. I'm giving the parameters to that people can recreate exactly the same curves in hi-res using Arve's tool, the adjust to taste/set-up. If you don't want to spend the time learning to use the tool, the solution is called Chad B (or any other competent calibrator)
. No one is stuck to anything, unless you're both out of time AND money (which rarely happens, but I acknowledge it can).
Originally Posted by fingersdlp
That is pretty cool. The black ones are pretty neat when you watch title cards fade in and out of black. I tried the hard clip setting on some earlier curves I did and it was interesting to see what was clipped. I also found there is more content than I would have guessed in the high range > 3500 for Fury Road. Not just explosions but glints off the exhaust pipes and such. I set it to revel any over 3500 and there was a good amount. Nice to have that so it is less guess work about what is getting altered. Just when I thought I might be close to the end experimenting. Oh well. Good stuff.
Why do you think I keep using (and recommending) to use Mad Max to check for highlights clipping below 4000nits?
It's one of the few 4000nits titles (along with Batman Vs. Superman) that actually uses the whole range. Many 4000nits titles (for example The Shallows) never go above 2500nits. Most never go above 1100nits.
Originally Posted by claw
I used the Work In Progress branch tonight and created two curves; one with 600 max brightness and the other 500. So far I am preferring the punch I get from the 500.
500 100 .020 4000 auto(215) .75 0 1
.020 resolved to bar 81 with both curves. Both resolve to 4000 nits was well.
It really wasn't difficult to use once I realized I had to load a default or other curve before doing anything else.
Great, please double check that your 500 curve does resolve down to 81, create another one resolving down to 68, and please do the testing requested above and in my initial post about setting brightness. We do need feedback on this to decide whether it's an acceptable practice for our projectors to resolve down to 81, so we only need one Unicurve, or if it does have a downside with some titles.
Originally Posted by IMDave
Hi Steve or anyone that is using the WIP Arve utility.
Don't know what I'm doing wrong, but no adjustments I make to "bb" seem to make any difference to the resolving brightness.
Using Ray's patterns Black clipping no.2, 64 and above all flash.
At bb=0, bb=0.05, bb=0.10, bb= ?? all flash.
JVC an Oppo set to 0/0 Br/Cr.
Manual iris, low lamp open.
Let the WIP do the auto calcs, with a max at 600: so, 600 - 100 - Auto ( 255 )- 4000 ( 0.75 / 0 / 1 )
I saw Manni01 previously mentioned that movies mastered at 0.005nits initially showed a raised black floor, which I'm also seeing, and that he noted that with the Linker in place to utilize the DI, that the DI was not closing down fully on FTB.
I've also noted that when a movie is loading, the DI closes to maybe 3mm wide, but as soon as there is a signal ( during sync ) the DI opens considerably. Is this correct?
Any advice gratefully appreciated.
The DI opening fully during sync is normal. It should, however, close fully when displaying a 0nits full black pattern. Don't check this with an actual title, you never know if black in a fade to black is 0nits, 0.005nits, or anything else, some have a raised black level even if the level is not raised otherwise, some are too short for the DI to have the time to fully close.
Please try to resolve down to level 81 in a separate curve and please compare both curves as requested above, once you have checked please report results back.
Note to all
: this testing isn't about which curve you prefer in your own setup, or whether my parameters work or not for you. I'm glad if they do, but they most likely won't, especially max brightness if you don't have 120-140nits peak white, and black brightness which will need to be adjusted using the Masciola patterns. I'm asking for specific feedback on a few different points in that post about setting brightness in HDR.