How to adjust a CMS system - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 2Likes
  • 1 Post By Kamikaze_Ice
  • 1 Post By Kamikaze_Ice
 
Thread Tools
post #1 of 11 Old 02-11-2017, 02:29 PM - Thread Starter
Senior Member
 
dschlic1's Avatar
 
Join Date: Aug 2011
Location: Tampabay, FL
Posts: 254
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 87 Post(s)
Liked: 29
How to adjust a CMS system

I have just completed the gray scale calibration on my LG OLED. I would like to try and adjust the CMS on this TV with HCFR but don't really know how to proceed. Each of the six colors has three controls: Tint, Saturation and Luminosity. What would the procedure be to adjust the CMS system on this TV?

LG OLED65B6P TV, Yamaha RX-V665 AV receiver, Sony BDP_S1700 Player, Klipsch F-30 speaker system
dschlic1 is offline  
Sponsored Links
Advertisement
 
post #2 of 11 Old 02-11-2017, 02:53 PM
AVS Forum Special Member
 
Join Date: Dec 2014
Location: Mississauga, ON, Canada
Posts: 2,090
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked: 327
Quote:
Originally Posted by dschlic1 View Post
I have just completed the gray scale calibration on my LG OLED. I would like to try and adjust the CMS on this TV with HCFR but don't really know how to proceed. Each of the six colors has three controls: Tint, Saturation and Luminosity. What would the procedure be to adjust the CMS system on this TV?
Tom Huffman's CMS Guide
Dominic Chan is online now  
post #3 of 11 Old 02-13-2017, 07:27 AM - Thread Starter
Senior Member
 
dschlic1's Avatar
 
Join Date: Aug 2011
Location: Tampabay, FL
Posts: 254
Mentioned: 10 Post(s)
Tagged: 0 Thread(s)
Quoted: 87 Post(s)
Liked: 29
I have read that guide and also several others. I am however somewhat dense and can't quite see my path going forward. For example on my set I have control over both the saturation and luminosity. Luminosity is directly measured via the Y value. I can see how to adjust it. However How do I measure the saturation? I know that saturation is the distance from the white point. I tried to adjust both the tint and the saturation using the target graph, but did not get any kind of consistent results. So I guess I am asking for some nuts and bolts assistance here.

LG OLED65B6P TV, Yamaha RX-V665 AV receiver, Sony BDP_S1700 Player, Klipsch F-30 speaker system
dschlic1 is offline  
 
post #4 of 11 Old 02-13-2017, 09:39 AM
AVS Forum Special Member
 
Join Date: Dec 2014
Location: Mississauga, ON, Canada
Posts: 2,090
Mentioned: 15 Post(s)
Tagged: 0 Thread(s)
Quoted: 1254 Post(s)
Liked: 327
Quote:
Originally Posted by dschlic1 View Post
I tried to adjust both the tint and the saturation using the target graph, but did not get any kind of consistent results. So I guess I am asking for some nuts and bolts assistance here.
You can watch the dE when you make the adjustments. It's also easier to display the "live" CIE Diagram which gives you a pictorial view of the effects of the saturation and tint adjustments. The diagram does not show the luminance, however.
Dominic Chan is online now  
post #5 of 11 Old 02-24-2017, 10:23 AM
AVS Forum Special Member
 
ConnecTEDDD's Avatar
 
Join Date: Sep 2010
Location: Athens, Greece
Posts: 4,584
Mentioned: 25 Post(s)
Tagged: 1 Thread(s)
Quoted: 1132 Post(s)
Liked: 1625
Quote:
Originally Posted by dschlic1 View Post
I have just completed the gray scale calibration on my LG OLED. I would like to try and adjust the CMS on this TV with HCFR but don't really know how to proceed. Each of the six colors has three controls: Tint, Saturation and Luminosity. What would the procedure be to adjust the CMS system on this TV?
Hi, these LG OLED's while they are amazing looking displays, their CMS controls are introducing problems when you are using them, or you can minimize the issues when you will do specific adjustments (not fully adjust all available CMS controls you have). If you look only dE charts, it's only a small percentage of success, if you look real content or verify using reference patterns then you will locate what kind of problems.

If you have my calibration disk then use that pattern:



See there: ***Official B/C/E/G6P OLED Calibration Thread

The way to receive the best possible picture from these OLED's is to use 3D LUT display characterization and calibrated thousand of color points without using LG's poor CMS internal controls, look here: HCFR - Open source projector and display calibration software and there: Xbox One S-UHD Blu-Ray Player Aug 2016

When you want to check or calibrate color gamut with internal controls of a display be sure that the software you are using has the correction settings for that and that you will use the correct patterns for this job.

Ted's LightSpace CMS Calibration Disk Free Version for Free Calibration Software: LightSpace DPS / CalMAN ColorChecker / HCFR
S/W: LightSpace CMS, SpaceMan ICC, SpaceMatch DCM, CalMAN 5, CalMAN RGB, ChromaPure, ControlCAL
V/P: eeColor 3D LUT Box - P/G: DVDO AVLab TPG
Meters: JETI Specbos 1211, Klein K-10A, i1PRO2, i1PRO, SpectraCAL C6, i1D3, C5
ConnecTEDDD is online now  
post #6 of 11 Old 03-04-2017, 11:52 AM
Advanced Member
 
harlekin's Avatar
 
Join Date: May 2007
Posts: 578
Mentioned: 56 Post(s)
Tagged: 0 Thread(s)
Quoted: 245 Post(s)
Liked: 72
My approach is the following (not weighing in on the "does adjusting the CMS introduce artefacting on the LG OLEDs" discussion. ) -

First make sure the Color Bars setting in HCFR is set to the one that represents what you can change with your CMS sliders. There are two different sets of color bars you can switch in between in HCFR.

Have a CIE triangle Window up at the same time you are adjusting Colors.

First I try to get the Color Luminanace close. (It's not a bar, its a percentage value.)
Second I try to get the point on target with Tint - using the Color Triangle as a reference.

If the point already is "close to target" - its mostly a matter of looking at the dE value and moving that colors Tint to the left and right (- / +) by increments of 2 - until about + and - 5 and see if the dE "jumps" in any direction. If it does you can eliminate the direction it didn't jump at (+ or -) and fine adjust.

At this point, you also have to adjust Luminance again - because it will change in accordance.

After Tint is dialed in, move to saturation.

Adjust saturation, adjust tint, adjust luminance.
-

The overarching guidelines are as follows:

Tint will move dE by the largest amount - but do dial it in, Luminance should at least be in the ballpark.
Once you hit the right tint spot, you'll have to adjust luminance again.
Saturation still moves dE but does so in smaller increments - so we dial in that after tint is set.
Once we've adjusted Saturation, we have to dial in Luminance again - Tint at that point might have to be fine adjusted - but usually it remains "on spot".

Also understand that if you dial in tint (especially true for tint - but holds true for the oder sliders in a lesser amount) at a colors 75% level - it might become "more off" at the 100% saturation level and so on -

To get "best" values on paper (color checker dEs) calibrate Colors at 75% saturation (HCFR setting bt709 75%), then switch to 100% (normal bt.709) - do a prim/sec color measurement run, do a saturation measurement run - then look at the saturation dE Graph - identify where what prim/sec jumps dE the most - then navigate to the data panel of that saturation sweep - pick the two saturation levels you try to mitigate between (lets say 75% and 100%) and do spot measurements (camara icon), while switching between them and correcting in the CMS.

(As a guideline - I was able to bring down 75% saturation dE to 0.4 or lower across the board - but then the 100% dE' might not "agrree".. )

That said - best values on paper not always correspond with best perceived results - so do several preset calibrations, and compare them over time..

Also - on the LG OLEDs - don't touch the global tint slider, imho. You'd get better skintones setting it to G4 for example (and THEN starting the CMS calibration) - but its always in sacrifice of another color sector (in that case afair cyan getting more off) - and skincolors arent that percetively wrong that you'd absolutely have to...

Even if you try to minimize Color Checker SG dEs (I have a few calibrations, where only two colors exceed dE3) - a calibration doesn't get "obviously better" - in fact you might not like the overall result.

If you want to experiment - try to reduce the "most correction needed according to measurments" colors corrections afterwards - while looking at images to see how the changes impact that color.

Make a habit of writing down your calibrations, although it is a hazzle because the TVs menu system is slow.

Always have a calibration preset at hand - where you didn't touch the CMS - and decide over time if you'd actually made an improvement.

Have fun.

h.

Last edited by harlekin; 03-05-2017 at 12:00 AM.
harlekin is online now  
post #7 of 11 Old 03-04-2017, 09:30 PM
Senior Member
 
Kamikaze_Ice's Avatar
 
Join Date: Jan 2007
Posts: 236
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 112 Post(s)
Liked: 58
Just adding my pennies.

Three options:
I) Don't touch the CMS controls for your OLED.
II) Make small adjustments only (or existing MINOR errors will become very exagerated at some points but not at others (not all gamut points are covered by CMS adjustments; coverage has some holes and hardly adjusted spots)
III) Find the CMS controls that are notably worse than the rest and never touch it but adjust others. (I've covered how _I_ check for this around page 30 in the oled calibration thread for these displays)

...
or if you want to include not using the controls as a way to "use" them :
IV) 3D LUT solution (The best solution for PC is also free; no need to buy a 3D LUT box for computers).

You could have just kept this post in the OLED calibration thread, y'know. We don't bite (well harlekin might if you have a 'sammich, I don't know lol.)

Also, T( )( )L made a post (HERE) where he has taken two pictures showing you what happened on his panel when making adjustments to yellow and cyan TINT by 5 steps (hint: color wrong on only HALF of the yellow/cyan coverage.
He was kind enough to check how his red looked at +30 saturation/luminance, which eventually led to me sharing what I did for red as mine was completely different than his with the same adjustments (this is the III option). Feel free to try the III option at your discretion.

The only real answer to the question is 3D LUT solution. The others are simply band-aids.
Personally I think the Bacon-looking band aids are the best, but it's not wrong if you don't agree with me.
ConnecTEDDD likes this.
Kamikaze_Ice is online now  
post #8 of 11 Old 03-04-2017, 11:54 PM
Advanced Member
 
harlekin's Avatar
 
Join Date: May 2007
Posts: 578
Mentioned: 56 Post(s)
Tagged: 0 Thread(s)
Quoted: 245 Post(s)
Liked: 72
Concerning the LG OLEDs - I currently follow a more radical approach - of make big changes - compare the results, and try to find preferences.

That said - I've never seen anything resembling those obvious artifacts on faces (skin colors) to begin with - but I have yet to look at a Color field and try to spot them to get primed what they look like..

(I've just croped and upscaled @ConnecTEDDD s color field using a Lanczos upscaling algo. so Ill look at that eventually. I will have to pause again for a few days until I am able to do so.)

(And yes the theory do then touch the "heavily adjusted colors and try to see the impact of them being "reigned in" stems from the idea - that you shouldn't make changes that are too big in the CMS - but thats a working theory. Im mostly looking at blue (heavy adjustments needed) - while getting rid of the magenta tint at the moment -- you could use the Casino Royale shot where Bond steps out of the plane in the bahamas - and look at the sky for example (use the establishing shot of the plane on the ground, just before.) Plane is white, you have greens and pastels in the image - its a good mix, but then you can use any material you are familiar with.)

Last edited by harlekin; 03-04-2017 at 11:58 PM.
harlekin is online now  
post #9 of 11 Old 03-05-2017, 03:15 AM
Senior Member
 
Kamikaze_Ice's Avatar
 
Join Date: Jan 2007
Posts: 236
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 112 Post(s)
Liked: 58
Quote:
Originally Posted by harlekin View Post
Concerning the LG OLEDs - I currently follow a more radical approach - of make big changes - compare the results, and try to find preferences.

That said - I've never seen anything resembling those obvious artifacts on faces (skin colors) to begin with - but I have yet to look at a Color field and try to spot them to get primed what they look like..

(I've just croped and upscaled @ConnecTEDDD s color field using a Lanczos upscaling algo. so Ill look at that eventually. I will have to pause again for a few days until I am able to do so.)

(And yes the theory do then touch the "heavily adjusted colors and try to see the impact of them being "reigned in" stems from the idea - that you shouldn't make changes that are too big in the CMS - but thats a working theory. Im mostly looking at blue (heavy adjustments needed) - while getting rid of the magenta tint at the moment -- you could use the Casino Royale shot where Bond steps out of the plane in the bahamas - and look at the sky for example (use the establishing shot of the plane on the ground, just before.) Plane is white, you have greens and pastels in the image - its a good mix, but then you can use any material you are familiar with.)
I don't think it's a good idea to use the image he used for this purpose. The pattern yes, but the picture probably is not the same thing (read: likely a slightly modified file for use on the web). And I'm not trying to speak for Mr. Aspiotis, but it wouldn't make business-sense to offer the patterns he worked on for free while selling them at the same time (ignoring the obvious difference; picture vs video vs disc). If they are the same it's still a nice collection covering a lot of ground and IMO worth the price. He can get a beer or 6 with that donation.

HCFR has this pattern anyways (Advanced > Gradients & Ramps > Granger Rainbow [hsl]), as well as others for this purpose, so just connect it to a PC and have at it. You have to use GDI and not MadTPG for this, because these patterns are apparently "animated" as far as HCFR is concerned which boggles the mind.
Personally I like the "horizontal" version (I might have this backwards...) so you see more points between black and white and less between hues.

I can't take pictures the majority of artifacts, but if you want to see what colors to focus on when checking for artifacts, simply set all the CMS controls to +30 saturation & luminance. Then look at a granger rainbow pattern.
The edge/rim of what each CMS control individually adjusts is where these changes will be most apparent.
Some skin tones in the right lighting are effected by this more than others.
Or even better, just play with the CMS controls while looking at this pattern

With only my panel as a reference, the vast majority of the kind of artifacts that becomes more apparent with CMS adjustments can easily be mistaken as an error in the source content (ala macroblock decoding). Some of this stuff won't be noticeable with patterns yet can see in content.
IMO, it's much easier to see with HDR content converted to SDR via MadVR. Errors caused by the HDR > SDR transformation will be very obvious (and will only happens to the highlights since HDR uses the fixed nit EOTF).
The reason I say this because HDR is simply more colorful in general.
The LG demos from HERE make for nice testing material for this purpose (using MadVR), and can be used for the same purpose for HDR mode adjustments.

Personally I'm the third of the three options in my last post.
Since red is the only color that is vastly different than the rest at +30 saturation/lumination, I simply don't touch it at all. I've not played around with this enough, but so far I've not had any problems by adjusting all but red as needed by measurements (obviously going the wrong way makes things worse).

I'm planing on seeing if using Color/Tint controls to adjust red and the rest adjusted with CMS actually works out. I know it works perfectly with test patterns, but I've not checked anything outside of CMS control point patterns (skin tones, 500 point color checker results, etc), which should help highlight limitations of the CMS controls.
That's a lot of points to check with my OCD process....


Sorry for yet another long post from me.
mo949 likes this.
Kamikaze_Ice is online now  
post #10 of 11 Old 03-05-2017, 05:36 AM
Advanced Member
 
harlekin's Avatar
 
Join Date: May 2007
Posts: 578
Mentioned: 56 Post(s)
Tagged: 0 Thread(s)
Quoted: 245 Post(s)
Liked: 72
I will use thw HCFR pattern then to evaluate. In fact Ill use both. Lanczos upscaling should do interpolation (?) between pixels so upscaling a gradation seemed to be thing that you could do.

Also the idea is that color gradation fields should trigger the artifacting, not that a "perfect" color gradation image will trigger...

So my counter would be that only graduation accuracy in this cas probably would not be needed to trigger the issue. But its a guess. Ill take a more accurate method to test for free gladly.
harlekin is online now  
post #11 of 11 Old Today, 06:50 PM
Senior Member
 
Kamikaze_Ice's Avatar
 
Join Date: Jan 2007
Posts: 236
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 112 Post(s)
Liked: 58
Quote:
Originally Posted by harlekin View Post
I will use thw HCFR pattern then to evaluate. In fact Ill use both. Lanczos upscaling should do interpolation (?) between pixels so upscaling a gradation seemed to be thing that you could do.

Also the idea is that color gradation fields should trigger the artifacting, not that a "perfect" color gradation image will trigger...

So my counter would be that only graduation accuracy in this cas probably would not be needed to trigger the issue. But its a guess. Ill take a more accurate method to test for free gladly.
If you want to really get into it, make your own gradients in Gimp/Photoshop/whatever and use ImageMagick on them.
Kamikaze_Ice is online now  
Sponsored Links
Advertisement
 
Reply Display Calibration



Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off