I finally had a chance last night to finely tweak my gamma, grayscale and CMS. I was very pleased with the measurements.
However I did not finish up until very late, and therefore did not have much of a chance to evaluate the results with real-world viewing. Therefore I will preface this report by saying that these results and settings should be consider experimental until I've had a chance to spend some time viewing with these settings and report back over the next couple of days. Edit: on May 30th: I have now spent considerable time with these settings and they checked out great, so these are my final CMS settings (until the next time
This calibration was done with an Spyder 2 trained to an EyeOne. With my previous calibration I had retrained the meter several months beforehand. This time I retrained it at the start and the conversion matrix from the training was materially different (about a dE variation of 3-ish between the two). Considering how long it had been since previously training the meter I'm going to trust these more recent results unless my subjective viewing shows otherwise.
It is important to note that this training takes into account any color shift that my Firehawk may (or may not) have on the image. Also this was done with the iris fully open and in Normal lamp mode. Others have reported that iris and lamp can have a material impact on the settings. Also note that the dE of 4 at 50% reflected in the attached .chc was corrected down to a dE of 1.6 but I did not have the time to rerun the grayscale included. My lamp has about 300 hours on it.
Also note that I made EXTENSIVE changes to the custom gamma settings in order to flatten the gamma at 2.2 (with the exception of a sharp drop at 95%). This took over 6 hours in one session followed by about 2 houes of touch-up work last night. Without these changes it was not possible to truly flatten the grayscale in this manner nor was it possible for me to get a flat gamma across the entire range (I have posted questions pages ago about this and the need to fix it when I didn't have to previously with a new lamp, but disappointingly never got a response from the experts here).
Here are my latest CMS settings:
r 2 -26 10
y 5 -48 38
g 0 -44 33
c 4 -49 39
b 0 -24 10
m -1 -30 15
For quick reference dE measurements at 75% are:
r 1.7, g 0.8, b 1.9, y 0.9, c 1.3, m 1.5
For quick reference dE measurements at 100% are:
r 1.2, g 1.6, b 2.9, y 3.8, c 3.1, m 0.9
Perhaps the biggest difference between these numbers and my previous results (also with the latest firmware) is that I did not find it necessary to set the green hue at -10 this time around. Others never found this necessary either. I'm not sure if it is the meter training or something else has helped here.
Attached are two .chc files. Final.chc is my latest. The CMS was calibrated using 75% patterns and the gamut measurements in this file were also with 75% patterns. The final_100percent.chc file shows how the gamut measured with 100% stimulus patterns. Note that at 100% cyan, green and yellow become a tag under saturated. This is relatively insignificant IMO.
I could likely increase the saturation a few clicks for these colors so that they would be slightly over saturated at 75% and on target at 100%. However I have learned through all my work and experimentation and the feedback of others that it seems best to give priority to 75%, so I'm leaving this as is unless I discover a good reason to do something about it.
Also note that a lot of this is really splitting hairs. For instance, when dialing things in I might have cyan at say a dE of 2.0 and spend an hour dialing it in to 1.0, but when A/Bing the before and after versions of the colors they are virtually indistinguishable. At any rate I like to get it as close as possible to hopefully provide a bit more calibration buffer for errors as the bulb ages and the calibration drifts.
Overall I consider the CMS to work incredibly well with the new software. There does seem to be some IMO immaterial difference between 75% and 100% but this does not bother me in the least. Before there was a huge issue because it destroyed some colors like red due to the linearity issue, but with this all resolved and behind us I am a happy camper.
Note that to get the best results with these settings you will need to have your grayscale as flat as possible. This assumes that settings are transferable from unit to unit, which seems to be the case with some units but not others.
Edit: Updated to remove experimental status and change to final, after now having spent considerable time reviewing real-world materials. Looks nice.
rs20_cms.zip 5.0810546875k . file