Originally Posted by atabea
I re-ran my auto-cal last night and I think the results are infinitely better than my first try. In fact, I dare say it’s about the best image I have ever been able to achieve. Having done it a second time, and with Manni’s patient coaching, I now realize how terribly wrong I did it the first time.
My gamma no longer has that weird “S” curve. I can now watch Deadpool (HDR) without that sub-burnt look that I previously observed in some scenes. And while Black Crush is greatly diminished, I can still see it to some extent (in Deadpool anyway and practically non-existent on Revenant. Don’t know what it is about the the Revenant but that movie looks the best of all.
I started off running a full gamma and color calibration in NATURAL Picture Mode, Standard Color Profile, 6500K Pre-set and Normal Gamma. All processing turned off, Low Lamp, Iris manual and fully open. After saving the results, I ran a color-only calibration for the various Iris settings (-5, -7, -9, -11, -13). I also ran an identical color-only calibration in USER 2 for Manni’s REC709 NF custom profile, at the same Iris intervals.
I then ran a full gamma and color calibration for SDR BT2020 in USER 3, REFERENCE Color Profile, 6500 K pre-set, still in Low Lamp, followed by a color- only calibration for USER 3, BT2020 Color Profile, at the various Iris settings.
For Natural, REC 709 NF, and SDR BT2020, I am using a custom 2.4 Gamma correction with +2 on dark gamma for a BT1886 curve.
After completing the above, I then switched to HIGH LAMP mode, Iris fully open, and ran a full Gamma and Color Calibration with User1, REFERENCE Color Profile, Normal Gamma, Manual Iris, all processing turned off. After saving I ran a color-only calibration of User 1, JVC’s BT2020 Custom Color Profile, Manual Iris full open, then at -5 and then at -7 etc.
When I was done with the JVC software/autocal/Spyder 5, I switched over to Chromapure/iD3 Pro and calibrated 100% white. As you can see in the attached images, the DE was around 3 so it didn’t take much to tweak it down to .3 DE.
Can't say for sure if my Spyder is a good sample so I will leave that to Manni to pronounce it either way. Like I said, my image is spectacular so it can’t be all bad. The only thing I would want to improve right now is to correct the small bit of occasional black crush----which I can only detect in Deadpool so maybe it's just that movie.
I am attaching some screen shots for the usual suspects to review. There are two things left for me to do:
(1) Figure out how to integrate my HD FURY so I can try SDR BT2020 and
(2) Figure out how to use the Sony Test Patterns and determine where to clip white in HDR.
Kevin said it’s pretty simple using the apple app so we shall see.
A hearty thank-you to Manni and all the other helpful folks on these forums. I think a lot of us Novices would have given up on our BCMs if it weren’t for such folks.
You're welcome, and I'm glad you got results you're happy with.
A few comments:
Your Spyder seems to be pretty accurate as far as 100% white is concerned. A dE of 3 at 100% white compared to the i1d3 is pretty good, although there is no way to know how accurate the i1d3 is, could be less, could be more than that, impossible to know without a reference meter. However, the i1d3s are usually very close so it's in the ballpark.
Your rec-709NF profile in custom 2 measures a bit undersaturated on the green-red side. Again it's impossible to know if it's the meter or the calibration. It shouldn't be bothering you as it's very minor, and is probably less visible than the slight oversaturation (and likely worse tracking) in standard.
Re using the Integral to get BT2020, it should be as simple as setting it to custom mode 10 and setting the JVC to your SDR BT2020 user mode (at least with the Panny, no idea with the Sammy), but if you have any issues post in the main thread or in the HD Fury Integral thread for help.
Re iris positions, I've stopped to do all of these because in my opinion the improvement doesn't justify the time spent. Plus you are wasting time recalibrating color for iris settings whihc are too close to each other, for example -5 and -7, likely overriding one with the other. Harry identified four ranges, 0, -4, -8, 12 and I used to us 0, -5, -10, -15. Neither are fool proof on the new JVCs, because the biggest difference is between -3 and -14, so whether you do -15 or -12 you won't be accurate if you move the iris above or below -13. I've given up on this. I just estimate the iris setting / lamp mode I need in each mode (-3 low lamp in SDR rec 709 or SDR bt2020, -7 high lamp in SDR) and I only autocal that setting. I'll redo an autocal when I need to open the iris in a few hundred hours, or every 500 hours or so. This change is detailed in post 9 & 10. Of course nothing should stop you from running a color autocal for each range, but at least only do it once per actual range (which you will have to identify on the new models), not so close that you override the last one.
I did try to look for ranges last time and apart from the big jump between -14 and -13 (which is when the dual iris kicks out for the first time I think), I couldn't find any consistent boundary, and the errors where quite low anyway, I doubt they would be visible. That's why I gave up doing this. which I don't find sustainable now that we have not one (SDR rec-709) but three modes (SDR Rec-709, SDR BT2020, HDR BT2020) to calibrate.
Anyway, good job!