| Forums | Register | Polls | Search | Statistics |
 (?)  
You must be logged in to post content on this forum.
Display Calibration Light Illusion Forums / Display Calibration /  
 

Creating LUT for Lumagen to calibrate a display

 
 
Page  Page 1 of 2:  1  2  Next »

Author nathan
ZRO
#1 | Posted: 6 Jan 2025 06:43 
A few more rookie questions, now that I have had a chance to activate the rental license to give it a test drive for a couple of days. I'm trying to create a LUT that will act as a calibration (inside a Lumagen 4242) for my projector (which is set to bypass all of its internal settings).

When Profiling, I assume the target should be UHD2020 since I want to display content that is created and delivered in that space.

For Probe Options, until I solve the above question of how to use my profile data for my i1D3, I'm just using the generic correlation file from Lightillusion.

I'm using the Characterization window to run through a sequence of patches generated by the Lumagen. I ran both Cube and Grey Ramp. I am unclear whether this is enough to create the LUTs I need for the Lumagen.

Then in the LUT generation, I chose UHD2020 at the source, and for my Destination, the measurements from the Characterization step and generated the LUT.

I went back to the Hardware screen in the Profiling area, and chose an empty CMS slot on the Lumagen, and uploaded the LUT.

But when I use the CMS slot that I have uploaded the LUT into on the Lumagen, it looks like I don't have a LUT there at all.

My guess is that I missed a step -- or made a misstep-- in generating the LUT, but I'm at a loss as to what it might be. Am I missing something obvious?

Author Steve

INF
Male
#2 | Posted: 6 Jan 2025 12:31 
This is really covered in the User Guided.
  • When profiling, the Target colour space is irrelevant, other than to provide something for the graphs to plot against
  • Yes, using the correct Corr. File would be correct.
  • You do not need to run a ramp and a cube profile, as the ramp is within the cube - but you can use different profiles that deliberately have larger ramp patches. That is all in the User Guides.
  • Seems you have not actually saved a LUT? Did you save with a unique name so you know what you are uploading?
  • (You can use Active LUT to test the LUT before uploading to verify)

Worth reading this guide specifically: https://lightillusion.com/3d_lut_guide.html
Along with the Lumagen guide: https://lightillusion.com/lumagen_manual.html

Steve
Steve Shaw
Mob Boss at Light Illusion

Author gordonF
ZRO
#3 | Posted: 6 Jan 2025 13:19 
As Steve has said. Did you save the LUT once it was created? How do you know there is no LUT actually in the CMS slot in the Lumagen? The obvious way to check if anything has been altered is to turn on service mode in scaler MENU 0 9 1 0 OK then go to MENU>OUT>CMS> CMS in use and go in to the 21 step greyscale and see if anything has been altered. If there is a LUT it will definitely have written new values in there.

Author nathan
ZRO
#4 | Posted: 6 Jan 2025 21:15 
gordonF:
Did you save the LUT once it was created? How do you know there is no LUT actually in the CMS slot in the Lumagen? The obvious way to check if anything has been altered is to turn on service mode in scaler MENU 0 9 1 0 OK then go to MENU>OUT>CMS> CMS in use and go in to the 21 step greyscale and see if anything has been altered. If there is a LUT it will definitely have written new values in there.

Thank you for pointing out this sanity check. This would have been a nice fix/mistake but yes, there are new values (nice 21 point values) written in the CMS into which I uploaded the LUT.

Steve:
You do not need to run a ramp and a cube profile, as the ramp is within the cube - but you can use different profiles that deliberately have larger ramp patches.

I double checked and apparently did not save the CUBE profile but only the RAMPS (specially the Grey Ramps Primary and Secondary 8 bits), so that is what I created the LUT with, and uploaded. This has a minimal amount of data for a 3D LUT and 1D LUT so in theory while not particularly nuanced and detailed, this should get me something reasonably accurate?

To create the LUT I selection Source UHD 2020, Destination, the Profile I measured, and made no other selections on the LUT Generation page. The resulting LUT said it was 42% within target, which makes sense. The display has a greater than 709 coverage but not even full P3 coverage. But this shouldn't matter? I simply want to have the content track correctly as far as the display is capable to displaying.

And for the TPG Range in the upload screen I selected SDR. Perhaps I should have selected HDR. The Lumagen manual page (https://lightillusion.com/lumagen_manual.html) doesn't discuss this but the one screen shot shows HDR selected. And perhaps when uploading a lut that maps 709 content to my measured profile I should select SDR for the TPG type in the LUT upload screen. Just in case, I did uploaded it again, using HDR in the TPG Range.

Then I measured again and those measurement are attached to this post. I ran these measurement to see if they are consistent with why I would expect with the LUTs in place, but they look poor.

Maybe low IRE readings are a challenge off the screen so perhaps I should do the 1d LUT profiling with the meter pointed at the projector (to make sure there is enough light to get a consistent result) and do the 3d LUT with the meter pointed at the screen (to take into account any color shift the screen introduces)?

EDIT: Some more investigation, and I believe these are the pieces I need to revisit.

1. I left Extra Delay at the default setting. I will try this procedure to set it, next time: "The Auto button can be used to set the Extra Delay automatically, by first opening a patch window on the display to be profiled, placing the probe on it, and pressing the Auto button."

2. The LUT I created was just using the "Grey Primary & Secondary Ramp +" measurements (and without the Extra Delay "auto" option set up). I was thinking this was enough to learn the workflow and get a basic LUT in place. I may have been wrong?

3. I did NOT do this in the Lumagen before the measurements that were used for the LUT: "Set the 2020 <-->601/709 = Auto Convert, except while calibrating. If the LUT is used to convert between 2020 and 601/709 then set the "2020<->601/709" option to "off". " Nor did I undo that after measuring for the LUT and uploading it, because I did not: "Once calibration has been completed enable the CMS1 → Colorspace → Colorspace = Auto Convert." So that might have impacted my results. But I am not sure whether my graphs indicate that.

4. And I did not do this:
"To check for the correct set-up the following tests can be used.
Using a controlled video source display a zero black image
Change the In/Out HDMI options in turn and check for accurate black levels
(A probe can be used to check black levels)
Generate a New Unity Bypass LUT, and use the LUT Tools to set Gamma to 2.0
(If the ColourSpace license you have doesn't have the LUT Tools option, use the LUT Image to make a similar gamma change via an external graphics program)
Using a controlled video source display the Light Illusion BrightnessCal image
Again, change the In/Out HDMI options in turn and test by uploading the Gamma Change LUT created above
When the correct Radiance configuration is set the Gamma LUT will operate correctly, altering the image Gamma, with NO change to zero black."

......because I don't see any in/out hdmi options in the ColourSpace manual and searching for "In/Out Hdmi" in the Lumagen manual also returns no hits.








Author nathan
ZRO
#5 | Posted: 9 Jan 2025 18:12 
nathan:
4. And I did not do this:
"To check for the correct set-up the following tests can be used.
Using a controlled video source display a zero black image
Change the In/Out HDMI options in turn and check for accurate black levels
(A probe can be used to check black levels)
Generate a New Unity Bypass LUT, and use the LUT Tools to set Gamma to 2.0
(If the ColourSpace license you have doesn't have the LUT Tools option, use the LUT Image to make a similar gamma change via an external graphics program)
Using a controlled video source display the Light Illusion BrightnessCal image
Again, change the In/Out HDMI options in turn and test by uploading the Gamma Change LUT created above
When the correct Radiance configuration is set the Gamma LUT will operate correctly, altering the image Gamma, with NO change to zero black."

......because I don't see any in/out hdmi options in the ColourSpace manual and searching for "In/Out Hdmi" in the Lumagen manual also returns no hits.

I suspect this (Change the In/Out HDMI options in turn) is a setting in the Lumagen, right?

Author Steve

INF
Male
#6 | Posted: 9 Jan 2025 18:46 
The graphs show issues with measurements.
Any time you see a point outside the CIE diagram you know something is wrong.
That may be Extra Delay, or something else entirly.

And the first two grey scale measurement are zero - so either clipping, or the probe just cant read them.
And you will likely need to use a large Cube profile, not just a Ramp Only, for that display.

And yes, the HDMI settings are Lumagen internal settings.

Steve
Steve Shaw
Mob Boss at Light Illusion

Author nathan
ZRO
#7 | Posted: 10 Jan 2025 06:17 
Interesting. Okay, taking into account those errors, I ran a similar set again. (Will do a large cube once I get this limited set successfully created.)

Still seems like on the low end the i1D3 is having trouble reading this JVC RS500 in bypass mode (no color profile loaded). This is the Color Munki Display version, reportedly capable from 0.003–1000 cd/m² which is just about as good for low light as the i1D3 family gets, IIRC.
















Author nathan
ZRO
#8 | Posted: 10 Jan 2025 06:19 
So I figured why not turn the probe toward the projector (with the diffusion lens on) and see if that helps with low IRE readings.

Still not great. Got a little better. Nice to see contrast off the lens is 55k:1. But maybe my workflow is still borked.
















Author Steve

INF
Male
#9 | Posted: 10 Jan 2025 12:57 
The RGB Balance and Clips graphs show you have an issue with the signal path settings.
You definitely have clipping in the near blacks, as the RGB Balance shows.
And something really strange is happening in the peaks, which may/may not also be clipping.

Likey a Legal vs. Data range issue in your workflow.

Steve
Steve Shaw
Mob Boss at Light Illusion

Author nathan
ZRO
#10 | Posted: 10 Jan 2025 18:24 
Steve:
Likey a Legal vs. Data range issue in your workflow.

Goodness, well, that's easy to fix! Looking better once I toggle that option:


















Author nathan
ZRO
#11 | Posted: 10 Jan 2025 18:28 
Any other settings look suspicious (based either on what you see them set as, or the results)?

Also, ran a small cube after the basic profile done and shown in these screen shots, and here it is, attached.













Cube size 7.bcs.pdf Attached file:
Cube size 7.bcs.pdf

 

Author titleexaminer92
ZRO
Male
#12 | Posted: 10 Jan 2025 21:22 
nathan

I'd suggest checking the Auto box on Profile Luma in the settings window when you run your SDR verification sequence to ensure your the measurements are relative to your actual peak white reading.

Based on your screenshots and .bcs file, your target luma is 100 nits while the actual reading is ~62.

Author nathan
ZRO
#13 | Posted: 11 Jan 2025 17:56 
titleexaminer92
titleexaminer92:
'd suggest checking the Auto box on Profile Luma in the settings window when you run your SDR verification sequence to ensure your the measurements are relative to your actual peak white reading.

Based on your screenshots and .bcs file, your target luma is 100 nits while the actual reading is ~62.

Thanks for the tip!

Will this simply make the charts / graphs / data more readable, or will this materially change the final LUT that I create and upload?

Author titleexaminer92
ZRO
Male
#14 | Posted: 11 Jan 2025 20:30 
nathan

Yes, it helps to make the graphs more decipherable. Ticking that setting on or off has no bearing on the characterization measurements used to create the LUT.

Author nathan
ZRO
#15 | Posted: 25 Jan 2025 04:38 
Sanity check, does this xy diagram show that my color points are all falling within 709? If so, either I didn't set ColourSpace to measure wider (though I really think with my Target Gamut setting, and my Hardware Options setting, I have asked ColourSpace to measure the 2020 space) or I set up my Lumagen (which I am using to generate the patches) incorrectly.

I suspect the latter. It's as if the Lumagen might be limiting the test patches it outputs to REC709?

But after reviewing https://lightillusion.com/lumagen_manual.html I don't see anything that might be forcing that restriction. Thoughts?

I'm thinking I must have incorrectly set up the BYPASS mode in the projector. The native gamut of the projector is between 709 and P3, and the most likely reason for the results I am getting is that I am still stuck in a 709 profile even though I thought I tuned off all the color profiles..... (this is a different projector from the one I was using to learn on, in prior posts).






Page  Page 1 of 2:  1  2  Next » 
You must be logged in to post content on this forum.
Display Calibration Light Illusion Forums / Display Calibration /
 Creating LUT for Lumagen to calibrate a display

 

 
 
Online now: Guests - 1
Members - 1 [ idm ]
Max. ever online: 192 [11 Jan 2023 08:39]
Guests - 192 / Members - 0