[PlasmaPZ80U]

So far, I tried two methods for color calibration on my Samsung B650 LCD: Blue Only Mode for Color/Tint with Auto Color Space and D2 for CMS with Custom Color Space. Earlier today, I decided to try setting Color/Tint with the D2 using Auto Color Space. I wanted to compare the results between the two methods on a level playing ground.

Blue Only Mode


Color 45


Tint G48/R52


D2


Color 44 (optimized for Y of measured red primary)


Tint G52/R48 (optimized for minimal overall xy errors of all three calculated secondaries)


Therefore, the D2 used Y targets and secondary xy targets optimized for the measured primaries (xy). Color was reduced by one click while tint moved four clicks more green and less red. The end result was much more natural looking skin tones that are a little less bright and less red. Before, the skintones looked too red and a little too bright. It seems the Blue Only method was okay for setting color but way off for setting tint. One click for color is a minor change but four clicks for tint is a major change.

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18213418


I tried two methods for color calibration on my Samsung B650 LCD: Blue Only Mode for Color/Tint with Auto Color Space and D2 for CMS with Custom Color Space.

I'm not familiar with the Samsung User Controls. Is the first option Blue Only Mode equivalent to a Blue filter and only COLOR/TINT controls? The second method matching the Rec 709 color gamut with full CMS controls?


How was the Red Y target choosen?


Can you post a CIE diagram of both methods with one as a Reference for direct comparison and a Y table for review?

 

[PlasmaPZ80U]



Quote:
Originally Posted by Plasma54321 /forum/post/18221174


I'm not familiar with the Samsung User Controls. Is the first option Blue Only Mode equivalent to a Blue filter and only COLOR/TINT controls? The second method matching the Rec 709 color gamut with full CMS controls?


How was the Red Y target choosen?


Can you post a CIE diagram of both methods with one as a Reference for direct comparison and a Y table for review?

The blue only mode is identical to a blue filter that doesn't leak at all. It allows you to set color/tint just like you would with a blue filter.


The second method for setting color/tint is setting color to the Y target for the red primary determined by the measured primaries (xy). Tint was set to minimize overall xy error of the secondaries (not just cyan), using the calculated secondaries as the reference. To be clear, both the target Y value for the red primary and target xy values for the secondaries were calculated from the xy of the measured primaries.

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18213418


two methods for color calibration on my Samsung B650 LCD: Blue Only Mode for Color/Tint with Auto Color Space and D2 for CMS with Custom Color Space. Earlier today, I decided to try setting Color/Tint with the D2 using Auto Color Space. I wanted to compare the results between the two methods on a level playing ground.

So the comparison is:


1) Blue Only Mode with Auto Color Space (default fixed gamut) with only Color and Tint controls to adjust.


2) Using i1 D2 meter with Auto Color Space (default fixed gamut) with only Color and Tint controls to adjust.


The HCFR data files above are a little confusing for comparison. The Secondary Color file only has YCM and is based on what White level. The "Color" file is for which of the cases (above). It would be nice to see a direct comparison on a single CIE Diagram with Y data tables and Color Y targets identified.

 

[PlasmaPZ80U]



Quote:
Originally Posted by Plasma54321 /forum/post/18229780


So the comparison is:


1) Blue Only Mode with Auto Color Space (default fixed gamut) with only Color and Tint controls to adjust.


2) Using i1 D2 meter with Auto Color Space (default fixed gamut) with only Color and Tint controls to adjust.


The HCFR data files above are a little confusing for comparison. The Secondary Color file only has YCM and is based on what White level. The "Color" file is for which of the cases (above). It would be nice to see a direct comparison on a single CIE Diagram with Y data tables and Color Y targets identified.

The second file is to be used as a reference measure to display the secondary targets for the measured primaries. There is no before file that shows readings for the blue only mode method.

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18231955


The second file is to be used as a reference measure to display the secondary targets for the measured primaries.

Finally got some time to review and understand your HCFR data comparison files. I have never used the Reference Measure like you did for the Secondary Targets. Good idea! I also see why you tried to minimize the xy errors. The xy/ref data row makes it easy to observe changes while making adjustments. Did you do this in real time?


I took your Excel data file and added some of Bruce Lindbloom's CIE Color Calculator results. Your original data and a Y % difference (yellow highlight) from Rec 709 row is in the first table (light blue).


The middle table (light green) is Bruce's RGB calculator results for each of the colors. The RED>255 values are highlighted as well as GREEN

 

[angryht]

^^ I think HCFR also calculates the R, G, B values. So, if you make your results, x,y,Y, editable, you can just adjust the Y values until R equals 255. Then try that as your target.

 

[PlasmaPZ80U]

I'm not familiar with this third method. I should note my contrast setting is maxed out at 100 and I've set super-white off on my PS3 to avoid seeing the color shift above 100% stimulus. Does this impact the third method? I don't wish to turn on super-white as doing so will force me to lower contrast and then compensate by raising the backlight (which will hurt my blacks and contrast ratio). Lowering contrast on my TV results in 1 fL drop per click. So, a contrast setting of 100 is 5 fL brighter than a contrast setting of 95.

 

[Plasma54321]



Quote:
Originally Posted by angryht /forum/post/18257482


... HCFR also calculates the R, G, B values.

The HCFR RGB values are not the same values as in Bruce's calculator.

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18257853


I'm not familiar with this third method.

The only controls that are adjusted are the TV's COLOR and TINT settings. All other controls are not touched.


1) Adjust COLOR to get Red Luminance Target = 14.773

2) Adjust TINT to get Yellow components RGB Levels of Red% = Green%


(for additional Details see Reference: Fifth Method for Setting Color/Tint Controls – Preventing RGB (>255) Red Clipping http://www.avsforum.com/avs-vb/showp...8&postcount=44

 

[PlasmaPZ80U]



Quote:
Originally Posted by Plasma54321 /forum/post/18258137


The only controls that are adjusted are the TV's COLOR and TINT settings. All other controls are not touched.


1) Adjust COLOR to get Red Luminance Target = 14.773

2) Adjust TINT to get Yellow components RGB Levels of Red% = Green%


(for additional Details see Reference: Fifth Method for Setting Color/Tint Controls – Preventing RGB (>255) Red Clipping http://www.avsforum.com/avs-vb/showp...8&postcount=44

I'll give this method a try when I get a chance. I'd like to first make sure I throughly evaluate my current settings (Color 44, Tint G+2) subjectively with plenty of BD movies.


I can definitely say the custom color gamut approach for setting color/tint is much better than the blue only mode w/ flashing color bars approach. Skintones look great largely thanks to the xy errors of yellow and magenta being quite low. Does magenta have any significance with skintones? I think it might in scenes where there is blue light projected onto skin/faces (like a night, evening, or stormy scene).

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18258195


Does magenta have any significance with skintones? I think it might in scenes where there is blue light projected onto skin/faces (like a night, evening, or stormy scene).

It may, but what I've experienced is that when Yellow's Hue is off just a little it greatly affects the skin tones. Also, when I focused on Magenta errors I did not see PQ improvements. Our eyes must tolerate Magenta color errors more than Yellow.


One other point with only two controls - COLOR and TINT the first is for setting Red Luminance that only leaves one control for the Secondary Colors. Therefore, only one of them can be minimized with a single TINT control. So I chose Yellow as being most important Secondary and having the most affect on PQ.

 

[PlasmaPZ80U]

I was looking at the Bruce Lindbloom method, which indicates the yellow secondary has too much red relative to green in it. However, the CIE diagram shows the opposite when referenced to the calculated xy target based on my measured primaries (xy). Why is this the case?

 

[angryht]

I found a similar result. In the end, I decided to use the Lindbloom calculator to determine the Y value of red, limiting the max to 255, then I used the other methods (Accupel color calculator or my spreadsheet) to determine the values of the targets (x,y) for the secondaries. I'm still evaluating it, but so far I'm happy with it and skin tones look realistic.

 

[PlasmaPZ80U]



Quote:
Originally Posted by angryht /forum/post/18315523


I found a similar result. In the end, I decided to use the Lindbloom calculator to determine the Y value of red, limiting the max to 255, then I used the other methods (Accupel color calculator or my spreadsheet) to determine the values of the targets (x,y) for the secondaries. I'm still evaluating it, but so far I'm happy with it and skin tones look realistic.

So, you used the Lindbloom calculator for color and your spreadsheet for tint?

 

[angryht]



Quote:
Originally Posted by Plasma54321 /forum/post/18258068


The HCFR RGB values are not the same values as in Bruce's calculator.

If you make the values editable, you can adjust the Y value of red. When the R value, down in the Data Table below, matches the Y of White (measured after the secondary measurements), the result is the same as what you get when you do the Lindbloom iteration. You just have to divide the Y for red by the Y for white to get the percentage.

 

[angryht]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18315554


So, you used the Lindbloom calculator for color and your spreadsheet for tint?

Correct.

 

[Plasma54321]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18315363


I was looking at the Bruce Lindbloom method, which indicates the yellow secondary has too much red relative to green in it. However, the CIE diagram shows the opposite when referenced to the calculated xy target based on my measured primaries (xy). Why is this the case?

The Yellow Hue = 0% (HCFR RGB Level => Red% = Green%) is referencing the Rec 709 standard. Your calculated secondary colors from your custom gamut is not the Rec 709 references.

 

[PlasmaPZ80U]



Quote:
Originally Posted by Plasma54321 /forum/post/18315887


The Yellow Hue = 0% (HCFR RGB Level => Red% = Green%) is referencing the Rec 709 standard. Your calculated secondary colors from your custom gamut is not the Rec 709 references.

In that case, wouldn't it make more sense to use the custom gamut target over one designed for a gamut my display doesn't have exactly?

 

[angryht]



Quote:
Originally Posted by PlasmaPZ80U /forum/post/18315931


In that case, wouldn't it make more sense to use the custom gamut target over one designed for a gamut my display doesn't have exactly?

It does to me.