AVS Forum banner

eeColor + CalMAN Enthusiast Calibration Questions

8K views 95 replies 15 participants last post by  sillysally 
#1 ·
Hey guys,

Yet another thread from me as I learn more and more. Having had some advanced time with my LG EF9500 and CalMAN Enthusiast yesterday, I got around to doing Saturation and Luminance Sweeps, along with the Color Checker patterns. Which got me playing around with the LG's CMS controls. And as I am aware, it has always been advised to just avoid the CMS controls for the LG OLED's as they are not very "accurate". But the OCD in me wants to correct the colors to get everything "perfect".

So I did some reading and found that the eeColor box would basically fit the bill as I do not watch any type of 3D movies or content...though I obviously cannot use it with my Sammy UHD player since it doesn't support 4k. As far as I can tell, I would just connect the HDMI out from my AVR to the eeColor, than connect it to my Display (I would use the secondary HDMI out on my Anthem receiver for the Sammy UHD player to bypass the eeColor). Than when using CalMAN, I would connect the eeColor to my PC via USB cable and after a few settings, CalMAN would take care of the 3D lut creation/calibration automatically, like a DDC TV/Monitor basically.

1) Do I do a full 20pt Grayscale Calibration like normal before creating the 3D lut?
2) I should leave all CMS controls/color controls set to default (0) and let the eeColor handle it all?
3) Right now all I have is an unprofiled iD3 (will be getting it profiled by a calibrator with his Jeti sometime in the future)...this would still be an acceptable meter to use for the 3D lut creation?
4) I am using MadTPG as my pattern generator with CalMAN...will this setup still work for the eeColor/3D lut creation?
5) Once the 3D lut is created, if I ever go back to touch up grayscale/brightness/contrast on the TV, do I need to redo the 3D lut/recheck it for errors?
6) Any severe HDMI handshake issues I could potentially run into using it with my Anthem, Xbox One, Mede8er, or DirecTV cable box (again, no 3D content at all being used)?

Thanks guys. I know I have been asking a lot lately but I find this a really fun part of the overall A/V hobby and am just trying to maximize my results with as little investment as possible (the eeColor is about the most I can spend for a 3D lut device).
 
See less See more
#2 ·
Hey guys,

1) Do I do a full 20pt Grayscale Calibration like normal before creating the 3D lut?
No, just calibrate 30%/100% white to D65 (or your preferred white point) using the 2pt. controls.
2) I should leave all CMS controls/color controls set to default (0) and let the eeColor handle it all?
Leave CMS at default and use the largest gamut preset available (sometimes called native).
3) Right now all I have is an unprofiled iD3 (will be getting it profiled by a calibrator with his Jeti sometime in the future)...this would still be an acceptable meter to use for the 3D lut creation?
yes
4) I am using MadTPG as my pattern generator with CalMAN...will this setup still work for the eeColor/3D lut creation?
yes
5) Once the 3D lut is created, if I ever go back to touch up grayscale/brightness/contrast on the TV, do I need to redo the 3D lut/recheck it for errors?
The 3dlut depends on the display maintaining it's as-profiled state, no tweaking necessary. If you do change something you may or may not have to re-profile. Most of the time the grayscale will be a bit bumpier than what you can get manually and you will be tempted to smooth that out. That's ok but you'll just shift these small errors somewhere else so there is nothing to gain by doing so.
6) Any severe HDMI handshake issues I could potentially run into using it with my Anthem, Xbox One, Mede8er, or DirecTV cable box (again, no 3D content at all being used)?
The eeColor is slow to respond to handshakes but once locked-in it's fine. Be careful that you select the correct memory slot (1-6) in CalMAN DDC corresponding to the selected slot on the eeColor.
 
#3 ·
No, just calibrate 30%/100% white to D65 (or your preferred white point) using the 2pt. controls.
Leave CMS at default and use the largest gamut preset available (sometimes called native).
Some follow-up/clarification questions:
1) I guess I am that big of a newbie still but...the eeColor/3D lut generated corrects for grayscale/gamma also? For some reason I always thought it only handled colors :eek:
2) The LG has "Color Gamut" for which there is "Standard" and "Wide". Using Wide shows obvious overblown (over saturated) colors. But when used in conjunction with a 3D lut, the Wide setting will end up giving better/more accurate results?
3) Can I make a custom Gamma curve for the 3D lut to follow? Right now I am using 2.25 but I have 1-5% ramping out of black, so 5% is around 2.15 and 10% is about 2.23ish due to the nature of the LG OLED's. I am still learning all the fine details of CalMAN but I couldn't figure out how to actually make a custom curve that would ramp out of black as described...

Assuming the above questions are correct, literally all I have to do is make coarse 2pt corrections to the grayscale beforehand, leave the 20pt and CMS controls all set to 0, setup the eeColor, connect it to CalMAN, use the 3D lut workflow, and the rest is automated? Once completed sit back and enjoy?

Thanks for all the help. Really appreciate it!
 
#6 ·
BCJ, why don't you just use madVR and DisplayCAL to create your 3D LUT? You're already using your PC to calibrate and watch. I don't see any reason for you to spend money on a LUT box that is not even 4K compatible.
 
#7 ·
I don't use my PC to watch any content. Everything I watch is a stand alone player (Xbox One, DirecTV, and Mede8er). The PC is just there/part of the chain if I ever did decide to start streaming or play games once I upgrade my graphics card in the Fall.
 
#11 ·
For CalMAN Ultimate users, they can use the Virtual Cube and export whatever file type they want. (It's not available in CalMAN Enthusiast.)

But since you have CalMAN Enthusiast you can do it ala 'Virtual Cube' (without having eeColor there) by selecting MadVR from CalMAN Direct Display Control list in CalMAN Enthusiast.

MadVR is using eeColor 65-Point Cube format also, so you have to select as file type the eeColor - MadVR TXT

You can follow these setup instructions here: http://www.spectracal.com/Documents/...QuickStart.pdf
 
#16 ·
Generally choosing an amount of points (10/17/21p) that coincides/compliments with the capabilities of your LMD and DUT.

Using time based IRP method has its value too when having control over how long (20/45/90min) is more of a priority (t-shoot/evaluate) rather then a standard industry accepted size as in 17/21p cLUT.

I recommend point based to start with. Particularly 10p cLUT.

One of the most valuable things that you will need to acquire is a critical/trained eye for seeing both the negative and positive picture details in the image and video. This will surely take time, experience, study, and observance. Begin by acknowledging and appreciating what a 10p cLUT is doing, seeing the improvements with your own eyes on your particular display before moving on to larger (14/17/21p) LUTs.
 
#17 ·
Forgive me if this seems like a simple question I should know but how many actual individual points are in a 10p lut? Is that meant to read as 10 to the power of three meaning 1000 total points? So 14p is 2744 points, etc?

I don't have what I would consider professional eyes in detecting the most minute +/- but I can definitely detect a significant amount of difference. Enough that my OCD is never satisfied hence finally upgrading to a 3D lut box.
 
#18 ·
Yes the numbers (10/14/17/21) are cubed, e.g. 10^3 = 1000 points.

Most people don't start off having a trained eye for quality characteristics of image and video. It is acquired just like any talent. The more you understand and know what to look for the happier you will be with your new hobby. So don't be too quick to dismiss the corrections being made as being not beneficial. Observe the before/after graphs showing corrections and try to visualize and find these differences in your images and video. The struggle for the untrained eye you will face first is that once you make corrections to your display you will not have a side-by-side comparison to view. The natural tendency is to adapt and accept the material being viewed as sufficient even when it may not be anywhere near a known standard. That is why calibration to many is not as important as it should be. You will have to rely on memory and your developing analytical abilities.
 
#21 ·
Thanks for all the help guys. I set color gamut to wide, cleared all the CMS and 20pt controls, adjusted the 2pt controls (high was all that needed fixing luckily so I didn't have to mess with the low end IRE), and CalMAN is in the process of making my Lut as we speak. I set it for 17x17x17 points.

Once I have the lut, if I ever want to go back and change the gamma curve will CalMAN recalculate the necessary corrections off of the Lut file already made or do I have to redo the whole thing?

Sent from my SM-G920V using Tapatalk
 
#26 ·
In addition to Chad's suggestion when you want the display to come out of black more quickly than a power law and your black measures 0, you will have to enter a user set black level above 0. Try checking BT.1886 EOTF and a black level of 0.01 cd/m^2.
 
#27 ·
Thanks guys.

Yea, I forgot to include in my post that CalMAN's full field pattern insertion was on (50% IRE every 60s). Plus, I have ABSL disabled in the Service Menu.
I have some free time tonight so I am going to load the profile in CalMAN and play around with it...try to get myself acquainted with all the subtle changes and settings.

Assuming the profile is correct/the reads were accurate and the grayscale measured correctly, I should be able to fix the grayscale/gamma issue without re-profiling? Or should I run a "quick profile" and see if it comes out correctly this time?
 
#28 · (Edited)
Don't get discouraged, something went very wrong. Probably in how you setup your madVR and HTPC, lets see how it looks when you get eecolor.

As Buzz or 10K pointed out, your Green acts like a Gamma control. So when you start lowering Green (Red and Blue to balance) your Y value (lumanice) becomes to low (dim picture).

For now do as zoyd suggests, that should work. The only issue with setting a black point other than 0 is you will probably notice the picture looks a little dull.
Maybe for now its best to use the sweet spot settings as shown below.
Target point of about 2.35 power gamma and the 2.4 setting in the EF9500.
Set contrast at 77, brightness around 50, color 51, OLED light at 42.

What I do is set the RGB High (gains) and Low (cuts), first setting my 100% to within a dE2000 of .3 or lower. Then I set the low RGB setting, keeping a close eye on the 'Y' value and x,y values for all 21 points. The key to a good pre setting of the grayscale is so there isn't any big settings adjustment for any of the 21 points.

Advanced settings.
Then I go into 5 IRE and set it so it is stable and within limits, then do 10, 15 and 20. After that recheck your brightness with the flashing bars, you may have to lower or raise your brightness setting slightly.

The reason for what I do in the low end is because of the brightness control being as course as it is.

I don't usually use Calman for a 3DLUT, but because of 1D calibration Calman has for the Lumagen Pro, I use Calman for both 1D grayscale and 3DLUT for CMS. Below is a link to the last Calman profile/1D/3DLUT that was uploaded to the Lumagen Pro 4444. http://www.avsforum.com/forum/139-d...elease-notes-discussion-109.html#post44129818

btw, as you may know the eecolor box will not use the GS calibration that Calman made, so you can ship the grayscale autocal calibration.

ss
 
  • Like
Reactions: L30Z3N and 10k
#29 ·
btw, as you may know the eecolor box will not use the GS calibration that Calman made, so you can ship the grayscale autocal calibration.

ss
Sorry, I'm a little confused by this statement. I didn't do a grayscale autocal calibration, I just did the 3D lut profile. The screenshots are the grayscale measurements I took using the 3D Lut profile applied to my display via MadTPG.

Are you saying even with a 3D Lut on the eeColor, I still have to manually calibrate the grayscale?
 
#34 ·
The correct method is to adjust the least amount of display settings as possible prior to the cLUT creation. If you must make corrections due to display non-linearity or deficiency it needs to be done prior to 3DLUT creation not after the viewing the verification results. The gamma issues seen in the picture show pertinent issues other than minor GS adjustments.
 
#49 ·
Ran a Lightning Lut (100 reads) and this time the 3D Lut Profile created was a lot closer to my target gamma curve. However, there is still room for improvement, mainly room for my pre-profile technique to improve which I am working on.

Question: Say the gamma after the profiling is finished is good but not great...can I manually adjust the 20pt controls in the TV lightly to fine tune any large dips or peaks AFTER the 3D Lut is applied? Or will this throw off the entire profile?
 
#50 ·
Yes, improve your Pre-Calibration settings methods is a good idea.

The answer to your question about Post(3DLUT)Calibration adjustments to your internal display settings was answered I do believe by a few different people. Maybe this thread or a different one you posted in.

I am getting ready to Re-Calibrate my favorite TV so ... here attached is some example of my Plasma panel Pre-Cal settings. I have only adjusted RGB gain of R/B for 100%W. I figured to show you this cause your display is causing you doubt as to what to expect. Next I too will calibrate with a 3DLUT using eeColor ... only I use LightSpace for this task :D
 

Attachments

#55 · (Edited)
So, I am really curious have you performed a Lightning (5min.) LUT using this method?

I have Calman E. and the Lighting LUT workflow is included but gives error when you try to use it.

And, I have never tried to use it via the 3DLUT cube workflow because I do not have a professional display like FSI (AM/CM/DM) or Sony PVM/BVM quality level.

If you did successfully use Lightning (5min.) LUT then how did it work for your OLED panel. Can you show before/after pics?


:)
 

Attachments

#56 ·
So, I am really curious have you performed a Lightning (5min.) LUT using this method?

I have Calman E. and the Lighting LUT workflow is included but gives error when you try to use it.

And, I have never tried to use it via the 3DLUT cube workflow because I do not have a professional display like FSI (AM/CM/DM) or Sony PVM/BVM quality level.
If you did successfully use Lightning (5min.) LUT then how did it work for your OLED panel. Can you show before/after pics?

:)

I was able to run lighting LUTs workflow within Calman enthusiast on a plasma display within the last couple months. It did not hang up or anything. Ran till completion. Given the nonlinearity of my display, it didn't really help but did satisfy my curiosity. As the instructions say, its for displays that are pretty linear to start with.
 
#62 ·
I know probably nobody's reading this LOL, but here's something else:

I tried to do a new profile with Argyll/DisplayCAL with all things identical except many more readings (4K+).

Same LUT generation parameters, full of garbage, esp in dark colors. I don't know if it's a DisplayCAL issue or because it think it has the data to try finer/darker adjustments or what. Even if I change the parameters and generate the LUT without applying the calibration (since the display is calibrated), it still has problems.
 
#63 · (Edited)
Haven't used Argyll in some time but when I did I got great results using it.
Make sure you have ASBL turned off, or software hack.
Use your High and Low RGB values set correctly, don't bother with a full RGB/grayscale calibration. Pay close attention to your Y (big Y) values, key is luma.
With my EF9500 I have brightness set at 51, contrast 78 and OLED light set at 42.
Warm 1, wide gamut but normal gamut will work fine also.
Also your meter may have a problem with reading very low light output readings.

ss
 
#67 ·
ABL isn't much of a issue when using the proper contrast and OLED Light along with a small triplet window. However there is some drift for 100% white, unlike a plasma that drifts down the OLED drifts up.
Most calibration software has a feature that avg your light output and corrects for drifting up or down.

ASBL is a issue when your meter reads dark patches.
When a screen is displayed for more than about 80 seconds the screen dims to protect against image retention.

ss
 
#72 ·
Ok, thank you - you've given me a lot to go through so I might not get time until this weekend to work on it. I will update this thread if I have any useful results.

I had previously done a meticulous 20 point calibration, but SillySally said it wasn't necessary for him so for this calibration I had only done a two point.

I can send the files to SpectraCal for help, but even if they find a problem I'm sure it will be fixed if I'm lucky in some new version maybe months from now and since this is for personal use I'd like it working sooner.

Is the file format for the eecolor 3D luts available for public? I'm a DSP programmer and am thinking about trying my own code to smooth the corrections. This seems like the type of problem we deal with constantly in audio but in video the developers maybe aren't so familiar with it. (When processing audio in realtime you must always compromise with the distortion introduced by making sudden or dramatic changes. I also wonder if they're exploiting the large LUT in the eecolor by doing an expensive interpolation over the points they sampled or if they're just linear interpolating corrections into the slots. I couldn't fix that though without generating my own tables - but just wonder.
 
#73 ·
Ok, thank you - you've given me a lot to go through so I might not get time until this weekend to work on it. I will update this thread if I have any useful results.

I had previously done a meticulous 20 point calibration, but SillySally said it wasn't necessary for him so for this calibration I had only done a two point.
20-Point Grayscale pre-calibration doesn't required, only do 100% White Calibration using RGB-High controls only.

I can send the files to SpectraCal for help, but even if they find a problem I'm sure it will be fixed if I'm lucky in some new version maybe months from now and since this is for personal use I'd like it working sooner.
You can send your files to SpectraCAL, but think that the latest CalMAN official software release (not beta) released @ 6th July 2016.

Is the file format for the eecolor 3D luts available for public? I'm a DSP programmer and am thinking about trying my own code to smooth the corrections. This seems like the type of problem we deal with constantly in audio but in video the developers maybe aren't so familiar with it. (When processing audio in realtime you must always compromise with the distortion introduced by making sudden or dramatic changes. I also wonder if they're exploiting the large LUT in the eecolor by doing an expensive interpolation over the points they sampled or if they're just linear interpolating corrections into the slots. I couldn't fix that though without generating my own tables - but just wonder.
It's not that simple, they are not use simple tri-linear interpolations, smoothing only the areas you have problem doesn't mean that you calibrate them also. Have you tried CalMAN AutoCAL only once; which failed; or you have tried 2-3 times and always you get results with similar problems (at blue/cyan region)?
 
#76 ·
I messed with it a little, (goofing off work), and I think I may have found the problem!! :eek::eek::eek::eek:

I'll let you know when I complete a recalibration to prove...

...I think it's the meter correction file. It did indeed show that cyan was far too green and then way over compensated far too blue (in some levels). Then I noticed a difference in the white level with a different tool and so tried it without the OLED meter compensation and cyan looks normal.
 
#82 ·
About the OLED table which is coming with i1Display PRO, X-Rite has created that OLED table for RGB OLED displays like Sony Trimaster EL or FSI OLED Broadcast Monitors etc.

The LG OLED's are WRGB and their spectral response is different, you can fix that problem if you create a meter correction table using a spectro (i1PRO1/2 or meter).

Some users compared that OLED table with i1PRO spectroradiometer and found that the normal LCD mode of i1Display PRO is closer to the i1PRO readings, I have sold all my i1DisplayPRO's/C6 to make any test with my JETI and confirm that.

But when a meter profile is off, it's not possible to report a so different (clipping) at your Cyan's. You will have more info if you measure the native gamut of your display with 10-Point Luminance as I suggested above.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top