The latest release of ColourSpace includes a unique Grey Scale Plus profile.
When this profile is used a totally different colour engine is used during LUT Generation.

The whole process of profiling using the new Grey Scale Plus profile and the subsequent LUT Generation colour engine has been developed for multiple primary colour displays - which as yet do not exist.

But, LG WOLEDs are a likely potential application, due to the additional White pixel.

At this stage we have only added the one Grey Scale Plus profile, but we will release further such profiles as multiple primary display technology progresses.
And if we get positive feedback from WOLED users, we will accelerate the roll-out for WOLED applications regardless of multiple primary displays becoming available, or not.

Steve

We are very interested in customer experience of using the new Grey Scale Plus Quick Profile for WOLED calibration, vs. a traditional large volumetric patch set based profiling & calibration.

Based on the new colour engine algorithms the Grey Scale Plus profile enables, there should be marked improvement in WOLED calibration, at least vs. the traditional Grey Scale Only Quick Profile, but should also generate results that are close to those of a large volumetric patch set.

Please do post your results and feedback here.

Steve

We've had some initial feedback from users testing out the new Grey Scale Plus profile, which enables the new multi-primary colour engine.

The first feedback is from a LG WOLED calibration.
The feedback was simply:

"Perfect!"



As can be seen, as good as a large patch set profile based calibration, but with just 27 patches!

The next feedback is from a user that tested the Grey Scale Only on a Eizo, just to see how the results compared to a normal 17^3 profile.
The feedback was:

"I did a cal with the grey+ to see what happens with an Eizo. Really excellent response.
Basically, as good as a 17^3 profile but done in about 2 minutes. Verifying was the slow bit "

More in-depth info on the result will be added by the above testers later - but with such great results it was worth posting some advance info.

Steve

I know Leon @liberator72 will tell you more about that new feature asap. But in the meantime, as I've tested it yesterday afternoon, I can tell you that it's working quite good. It requires only 60 mintues of your time (30 minutes warm up, SM WP calibration and 15 minutes of 1000 points verification included) and it gave me these results on my LG OLED55C8PLA:



It has 2 colors slightly above the dE of 2.3 (one dark yellow and one dark brown) and it would need a gamma correction in the low light area. Easy tasks.

This is already faster, cleaner and more efficient than ANY dawn, aurora or manual calibration. If you're in a hurry or you just need to calibrate your brand new TV while it reaches its break-in running hours, this is the best option. Of course, the more points you measure, the better.

Following on from Miki's post above, I spent quite a bit of time testing this new feature over the weekend (sorry it's taken a while to leave feedback) and I concur with Miki's assessment.

In my case, I ran the new Grey Scale Plus sequence a total of 6 times to check the results were repeatable, which they were.

Tests were performed on an LG OLED65C9, starting from unity 3D LUT and peak 109% white pre-calibrated to my target white point, and panel native gamma response. Below are two screenshots from 1000 point verification profiles of two of the resulting calibrations, one targeting Rec709, D65, 2.4 gamma, and the other targeting Rec709 with an alternate white point and 2.4 gamma.

Rec709 D65


Rec709 AWP


As it happens, I had actually reprofiled my C9 on Friday evening with a full volumetric custom patch sequence, and using Miki's excellent spreadsheet, here is the summary of the results from that calibration.

1000 point verification of a volumetric display profile


The results of the Grey Scale Plus LUT when verified with 1000 points


Pretty impressive from just 27 measured points, and visual verification with test patterns and real content is just as impressive with no banding or other artefacts that I have been able to find as of yet.

Of course, a full volumetric profile gives you the confidence and certainty that you display will be calibrated to it fullest potential, but for a quick calibration (it takes literally a couple of minutes) it produces highly impressive results.

There have been a number of comments on these results elsewhere, and bears repeating here:

"So with this new method, your raw native panel profile has only 27 points in it. Then ColourSpace will generate the entire LUT from just those few points. With very good results.
Does not seem right; based on my experience with another popular automated calibration program where the program can make thousands of measurements and still get visual errors."

As well as:

"That's some serious black magic going on there. That's awesome!"

And from one of the above users who ran the first tests:

"Manufacturers should really look to this to calibrate their accurate picture modes. It's one minute job and I'm pretty sure it can be industrialised. Wake up manufactures!"

Whichever way you look at it, the calibration results of the new Grey Scale Plus profile, and new Multi-Primary Colour Engine, really are in a league of their own.
And we are now expanding the profiles that will work with the new Multi-Primary Colour Engine.

Steve

Here are my results on 65" CX.
1 hour warm up, Greyscale Plus profile, verification using the patch set from LG guides.

I really like that spreadsheet for doing further analysis on verification. Especially the error by nit range. Is that available to download?

Please, write an email to geniusonecal@gmail.com, I'll explain you all there. Thanks.

The latest ColourSpace release includes another update to the Multi-Primary Colour Engine capabilities, with the addition of a Grey Scale Plus Large profile, and the ability to use user generated .csv files of the same type.

The .csv files can have different grey scale patches, as well as different matching values for the RGBCMY patches.
(Lowering the value of the RGBCMY patches can be used for HDR profilin, for example.)

The next development will be the alteration of the existing Primary & Secondary profile to also enable the Multi-Primary Colour Engine during LUT generation.
(We will likely also change the pre-set profile names to better reflect their use and operation...)

Steve

Very very very good!

New info, and an example of the benefits of the new 'Plus' profiles, has been added to the Advanced Operation page of the website.

The LUT Preview images really show how the new Multi-Primary Colour Engine improves calibration results, even on a normal LCD display, with low Blue gamut, which would otherwise require a large volumetric profile to calibrate.

Steve

Grey Scale Only

Grey Scale Plus

Apologies if this is a silly question, but I'd rather ask and risk making a fool of myself than to not ask. How would you do those two easy tasks? For the gamma correction, would you use the 20pt greyscale adjustment in the user menus after the 3dlut has been loaded? Then to correct the 2 bad colours, without using the CMS, I wouldn't know what to do.


Agreed.

Hey Tickle, you just need to add the problematic points in your profile patch set if they're not already there or, only for gamma, you could use a parametric value.

Thanks!

I've had great results using this too. I've tested it on my LG OLED55C8 with peak 109% pre-calibrated to liberator72's alternative white point of x0.3120, y0.3356. This is on my ISF Bright "day" mode with which I was aiming for ~180 nits and 2.2 gamma.

I echo the earlier comments, it's great. An average dE2000 of 0.6548 is amazing for so little user effort; I see charts proudly posted elsewhere with "before" looking terrible and "after" charts' points all hovering below 3 but an average certainly well over 0.6!

I actually ran it more than once because I made some newbie mistakes which made previous LUTs invalid. The exact mistakes aren't relevant, however on the first two runs I had used "Grey Scale Plus Large". The time with this option to create the LUT was over 11 minutes on my poor little Intel Core i3 @ 2.1GHz CPU. It was quite disconcerting, so if it's not a fault on my part, I would welcome (if it's possible to predict that it's going to be slow, say >20 seconds) a progress bar showing whilst the LUT is being calculated to reassure the user. The third time, which is this one in the report attached, I used "Grey Scale Plus", fewer points, and it took only 40 seconds which was much faster, but i understand this is an order of magnitude slower than a newer spec PC. I'm assuming that to create a LUT from so few points involves quite a lot of "sums", so it's not a criticism just an observation!

Attached file: wide graphs report.pdf