Originally Posted by claw
I haven't figured out yet how to use the wip version so I created your 800 version with the master branch.
800 100 335 4000 0.75 0 1
Since a previous curve: 800 100 600 4000 0.3 0 1 was a little too dark for me, I decided to also create a 600 version as well before testing. I noticed that your soft clip values were around 41.875% of the brightness value so I calculated the 600 curve soft clip at 250.
600 100 250 4000 0.75 0 1
JVC and Oppo brightness and contrast at zero. Using the Masciola patterns, both clipped very close to 4000 nits. The bar below 4000 flashed a little fainter that the bars below. I think I also noticed a slightly raised black floor with these compared to the previous 800/600 curves one but I will wait until I can figure out the wip version of the tool to attempt to deal with that.
I watched portions of Lucy and two others I recently viewed, Gatsby and Gods of Egypt. I like to test with the Gatsby mansion party scene with all the costume jewelry, bright chandeliers, and fireworks in the dark night background. While the 800 version was watchable, I preferred my 600 for the bit more brightness and punch it brought.
I then watched portions of Jupiter Ascending and Arrival. Again I could watch Jupiter with the 800 curve, but preferred my 600. With Arrival, I needed my 600 version.
My situation is different than most since I am still projecting to a 118" 16:9 painted screen so my gain is likely negative. I have not measured it.
I won't use high lamp so this was with Low Lamp, zero, Auto 2.
Now to try to figure out how to use the wip version of the tool.
Plots of the two curves attached; 600 on the left and 800 on the right.
Update: I also created an even brighter curve:
500 100 210 4000 0.75 0 1
I just barely reach the 4000 nit bar with this one. But it is bright. Second plot attached.
Thanks for this. It's good to know you're resolving up to 4000nits without any change.
Just to clarify, I'm not really looking for an evaluation of these curves, they might or might not work for you.
Assuming you are using, as recommended, YCC 4:4:4 (or 4:2:2) 12bits in standard HDMI, what I'm interested in re feedback (wip is necessary, knowledge of metadata of titles recommended) is:
1) When setting brightness and contrast to 0,0 in both the JVC and the Player (Panny preferred as this is my player), which number do you need for black brightness to not crush black?
2) When resolving down to level 81 (77 black) instead of level 68 (64 black), do you see the improvement in the black floor for all titles mastered to 0.005nits (for example, Lucy)?
3) When resolving down to level 68 (64 black), do you see a noticeable increase in shadow detail for the remaining titles mastered to 0nits (for example, Hitman 47)?
I need this feedback to see how many people experience black crush with Arve's curves in standard, and how bad it is.
I also need this to see whether there is a potential downside re shadow details in using a Unicurve which resolves black down to 0.005nits instead of 0nits. I haven't found such a situation with actual content, so if someone finds one I'd like to look at it (title, chapter, and timecode please).
The only thing I notice is a significant improvement (as in noticeable) of the black floor for most titles, and no visible black crush for the remaining titles.
Please note that it is likely incorrect to do this, but I'm after the right effect, not after a by the book calibration that looks worse than it could