Originally Posted by RIppolito
I am trying to calibrate HDR on an LG 2016 OLED. When I do a grayscale measure, and display RGB...
- I see that as IRE increases from left to right, the numbers for RGB increase, as well. What do the numbers represent?
- I often hear advice to simply balance RGB for a given IRE. With three different numbers, how do I best balance them?
- do I try to get them as close as I can to the D65 dashed line?
- do I pick the one that will require the smallest adjustments to the other two, and move them?
- something else
Thank you, in advance...
Depends on 2pt or 10/11/20pt.
For 2pt (offset and gain, which sets can use other names for those), you are trying to balance RGB for the lowest dE WITHOUT gamma included.
For that, you have the 20/80, 30/80, 20/100, 30/100 patches paired to try to reduce it to fully balanced.
What I personally do, because how it is implemented on some sets, is to ONLY change blue and red at the high end, whether with 100 or 80% patterns, as changing green can effect contrast features on some sets and thereby effect the tone map a bit.
For the low end, some say do NOT change the offset because it can throw off your near black significantly! It can also throw off the overall tone map is argued as well.
Personally, what I like to do is to minimize what changes are done at offset, as it can raise black levels and does throw off for HDR anything around the 10% and below mark. But, because I focus on the middle first, then the low end, I do allow the lowest end to go a little further off.
I do full grayscale sweeps. Then do the high end (gain) values for RGB. Then, I do another sweep. I move only one offset color at a time, then do another sweep. Then repeat while trying to maintain low end without overly increasing the errors there (no more than 1.5 dE on 5%) while getting the middle closer. Then rebalance the gain. Continue doing sweeps while monitoring the overall dE average for the grayscale.
Now, that means I WILL have more errors at the near black than others here. Facts ARE facts. But I can, while using dE ICtCp with my set, get it to 0.48 dE across the grayscale before I ever begin the grayscale with gamma using the 10/11/20 point corrections. Most will tell you just let the low end be because you can still correct it out on the 10/11/20 pt with gamma corrections. Probably a better practice overall and what I would do if I was calibrating someone else's display.
In any case, with 10/11/20pt correction, you need to switch it in preferences to include gamma in the dE calculations. Then, you try to balance all three of them (except at 100% where you only change red and blue) to 100% while also matching the Y (luminance) value. By switching on to include gamma with dE, it will automatically include the luminance in the target value.
Now, you should try to adjust the near white slope target values to match whatever your display does. If it hard-clips, leave BT.2390 not selected. If it rolls off at the high-end, then select BT.2390. You may click override targets, then change the NW value and % value until the roll-off target roughly matches the natural curve at the near white for your set. Then you can start your gamma passes. Many like to sweep downward when doing gamma, starting at 100% and moving to 5% when doing gamma. It may take multiple passes to minimize the error values. For my own sets, I target getting under 0.3 dE avg across all of gray scale with gamma. Some sets just can never get that low (and some projectors for sure). If you can get to 0.22 dE with gamma, you are golden. In the 0.15-0.18 range, I have rarely achieved that with my sets, but if you can get it with yours, blessed. That is more than anyone really could ask. Technically, anything under 0.5 dE you could move on to CMS, but try for lower than 0.4 dE if you have the time and patience and your set or projector allows for it with tuning.