Basic Guide to Color Calibration using a CMS (updated and enhanced) - Page 50

One possibility is to create your own private HDTV cable channel with test patterns by using the ZvBox 150, and then use your STB to tune to that channel.

I have a Pioneer Elite 151 which I am having trouble getting the colors to look accurate. This is my second 151. Both were pro calibrated, but the second monitor has a yelow push to it which I am having trouble fixing. I have an iI Pro meter and ran a CMS plot using Calman which shows a large blue error. I have attached a file of the reading that I took.

I know the Pioneer CMS is limited to adjusting hue only. Would an external video processor eliminate this error?

Any help would be appreciated.

 

cms with color decoder set to 4.pdf 72.2490234375k . file

I have tried turning down the color decoder, but the colors still look incorrect and when I run the CMS on Calman there is still a large blue luminance error. I do not have your knowledge base regarding calibration. My understanding is that a video processor will only decrease color saturation (add white). Since it seems that I have blue that is undersaturated, a video processor may not work. However, since I have never used one I am not sure if there is anything else that can be done to correct the color problem on this monitor. Is there any type of external video processor that can correct the problem???

guys, the numbers I posted a few posts up are from a Samsung LN52B650.

I wanted to follow up on the post above, as I've now read about 1,000 replies in this thread.

Indeed, it appears that my issue at this point is an inability to further saturate my red and green. My Sammy--like all Sammy's, it appears-- has its primary colors set with opposings at 0. My Red doesn't seem to be able to get any closer to the target red. blue and green are already at their lowest settings. as such, I just went ahead and optimized for Y and just about left my x/y to fend for themselves. that resulted in pretty accurate scoring for every color except red and magenta, which are floating around a DeltaE of 11 for red and 8 for magenta.

so my question is this: considering I can't saturate as far as I know meaningfully at this point...and knowing my Y's are fairly accurate now, should I call it a day and be happy with this?

appropriate file with relevant information about TV model and settings added to the notes attached.

Please, any insight would be greatly appreciated. For peace of mind purposes if nothing else.

 

1.30.10.zip 4.6845703125k . file

The LG 55LH90 appears to be very calibration friendly.
Here are some recent results using the Chroma 5 and internal patterns for gray and AVSHD for color. LD was on but full screen patterns were used.
Although only a 2D CMS is available, adjusting the color control for L and the CMS for color and hue gets it pretty close.

While I normally use ChromaPure, Calman was used here to get smaller .PDF file. The .PDF files produced by ChromaPure average 2 MB and exceed the AVS file upload limit. Is there a way to get these smaller ?

 

LH90 V2.pdf 131.677734375k . file

 

LH90-CIE.pdf 156.7275390625k . file

I got datacolor spyder3 colorimeter and downloaded the HCFR.
I read GrayScale & Colour Calibration For Dummies and I have the DVE Basics BD,but the test patterns in the DVEHD Basics are just good for the grayscale...(IMO)
How can i burn the HCFR's test patterns in a dvd in order to fix the prim/secon colors ,etc. in my Panasonic 50G15...?
Thanks and regards,
eugene

Pick the appropriate file that your Bly-ray or HD player will be able to read, download the file and burn to a regular DVD disk using a DVD burner.

sorry i was a bit short on this... what i need is the HCFR's test patterns files in order to use the software with the spyder3 colorimeter....the DVE HD doesn't have it
thanks!

A CHCFR sequence can be found on the AVSHD image.

I'm not sure if this is the right thread to ask this question of if it's been addressed already. Please re-direct me if appropriate.

I'm using HCFR and an i1 display2 probe to calibrate my LG 47LH90 TV. I've done a grayscale balance and there is something about the gamma for each of the primaries that I don't get. I've attached a .zip containing the .chc that shows this.

My strategy is to first balance the primaries at 10 - 100 IRE. Then, using the Y at 100 IRE calculate the Y required at each 10 - 90 IRE to achieve the target gamma (2.22 in my case). I then adjust the primaries at each 10 - 90 IRE for balance and the calculated Y. The resulting average gamma looks great. But the gamma of the primaries, especially red and blue, depart from the target, especially at the higher stimulus levels. The luminance graph is consistent with what I'm seeing in the gamma graph.

As a specific example, the 80% stimulus is the best in terms of balance and is very close to the target Y. At 100%, Y = 250.693. A target gamma of 2.22 requires that Y at 80% be 60.93 * (the Y at 100%) = 152.76. I'm measuring Y = 152.665. Also, x = 0.313, and y = 0.329. HCFR presents the avearge gamma at 80% to be 2.22. However, red gamma = 2.36, green gamma = 2.18, and blue gamma = 2.19. Since the grayscale is balanced and hits the right Y for a gamma of 2.22 I would expect the primary gammas to all be very close to 2.22.

What am I not getting?

 

Color34.zip 4.375k . file

What you're missing is that although they're very good, your RGB balances aren't straight lines, they do vary a little, the other curves reflect this. The biggest issue I see for your grayscale is that you're running out of blue a little at the high end. Drop your peak white 1 click and go through the exercise again without raising blue at 100%. That should clean it up a little more.

Why do you think I'm running out of blue at the high end? If anything, it may look like I'm running out of red between 70% and 100%. Anyway, I haven't presented my settings so you couldn't see this, but, there is plenty of headroom left to adjust all of the primaries. The present settings are just the closest I could get without having a bigger error in the other direction.

One thing I just noticed. If you look at the luminance graph at 90% this shows the biggest discrepancy for red. Basically, red is about 3.5% lower than the green. Yet the measure panel/Gray scale shows them within 2%.

Well, I could be wrong, but it looks to me like blue's slop (on th eluminance graph) is lower between 90% and 100% than it is between 80% and 90%. If you're running out, that would also explain the lower gamma for blue in that range. Just because you have the range to raise the value of the control doesn't mean the display will respond to it. It's the display that's running out, not the setting.

Rowland,

This is how I understand it. The RGB graph is showing the mix (think saturation and not luminance) of colors at a certain stimulus level. The gamma graph is then showing you the luminance for that same stimulus level in relation to your end points (black and white). When you then break down the gamma graph and show the individual colors, it is then showing you the individual luminance values for each color in relation to that particular color's end points.

-- edit --

I went back this evening and crunched some numbers so I could remember the specifics of how this actually works in HCFR. I'll document how I believe it works here. An example is the easiest way to show how the individual RGB gamma numbers are arrived at. Say you have the following measurements:

100% Y = 100 cd/m^2
90% Y = 80 cd/m^2
0% Y = .10 cd/m^2
Blue % on RGB graph at 100% = 105.31%
Blue % on RGB graph at 90% = 100.45%
Blue % on RGB graph at 0% = 110.06%
HD709 is the reference color gamut chosen in the Preferences.

So, what is the display gamma (assuming black level compensation) of blue at 90% stimulus?

100% Blue Y = .072 * 100 * 1.0531 = 7.58232
90% Blue Y = .072 * 80 * 1.0045 = 5.78592
0% Blue Y = .072 * .10 * 1.1006 = 0.0079243

gamma = ln((5.78592 - .0079243) / (7.58232 - .0079243)) / ln(.90) = ~2.57

The reference color gamut chosen affects the results a bit as well - both the first multiplier (reference Y value) as well as the percentage of saturation (the mix of colors at any point) which is the second multiplier.

So, if you alter the color balance mix at the end points without changing anything else, you'll have an impact on the individual RGB gamma points on the gamma graph. After re-working through this I'm not sure how much value there is in looking at it - as Tom H. points out if you get each point exactly at 100% RGB balanced then each color will have exactly the same gamma. If you get each end point at 100% balanced then the individual gamma graphs for each color will basically show you the same thing that the RGB white balance graph shows - namely if at any point you have too little (too dim) or too much (too bright) of a color. Based on all this, I've removed some of my original post since I think it gives the impression that looking at these individual RGB gamma graphs may be useful. I think I've convinced myself that they're really not.

-- end of edit --

So, in your case how I would interpret this would be that while the overall luminance at 80% hits your 2.22 gamma target, that luminance is not "perfect" for each color according to HCFR's calculations. Red is a bit dim, while green and blue are a bit too bright. Anyhow, based on what I've now written above, I would not look at the RGB gamma graphs and as long as you concentrate on white balance, you should be fine.

hope this helps,


--tom

This seems to be exactly what's happening. I edited the measurements panel so that the 100% stimulus was more balanced (it had 103% red, 99% green, 99% blue). After that change the gamma and luminance look better. I guess there's 2 ways to look at what gamma is for a particular color.


Actually, the LG 47LH90 it does. The red, green, and blue can be set for each 10 - 100 IRE. The CMS is just 2 dimensional though. Kind of a warped 2d. There are 2 knobs, one changing "mostly" saturation, and the other changing "mostly" hue.

Rowland, I've revisited some of my calculations from awhile ago and have re-done my above reply. I've outlined how I believe HCFR is calculating the individual RGB gamma graphs and in the end, I don't think they're particularly useful to look at since all the meaningful data is in the RGB white balance graph/measures.

good luck with the LG,


--tom

Has anyone had success in using the CMS on this projector? I've been trying with a Spyder3 Enhanced and CalMAN and I can't seem to get things to work properly. I suspect it's the nature of the Spyder rather than the projector. But it might also be my lack of knowledge in how to go about doing a proper Color/Tint/Gamut run.

Is there a very detailed step-by-step document on this key area that might get me going? The help in CalMAN doesn't seem to flesh out enough for me, or at least I can't find out where I might be going wrong. How might the different changes to brightness, saturation, and hue affect each other, if at all? Should I try and get Y set first for primaries, after White, and then go for secondaries? Or should I do, say Red first for Y, and then hit Cyan, with follow-up for other primary/secondary compliments?

I have not used that projector but what controls are available in the CMS? In general, what I would recommend is setting your Color/Tint in the main menu so your color decoding is as correct as it can be given your primaries and white point (I'm assuming you've calibrated your grayscale and set contrast/brightness). After that I'd then use the CMS controls to optimize your color gamut so it matches as best it can Rec. 709. It sounds like the CMS has 3 separate controls for each color? If so, assuming they are working properly, saturation will move the color in/out from the white point, hue will rotate the color around the white point and brightness will affect the luminance of the color. I would first use saturation and hue to get the x,y values of each color as close to the standard as possible then I'd set the brightness for each color so that it minimizes error as compared to the standard using dE1994 as the error formula.

hope this helps,


--tom

It has a menu for RGB (gray-scale) and another RGBCMY for CMS (18 separate controls, primaries and secondaries with hue, saturation, and brightness for each of the six).

I've used a variety of DVD's including GetGray to visually zero in on brightness, contrast, color and hue.

Interesting that your recommendation is the reverse of the CalMAN program help instructions.

No matter. I think it's more likely the fault of the meter (Spyder3) than the CMS. Hopefully, once I get the ColorMunki in house I'll be able to validate that, or maybe pin the blame on the Epson's CMS.

Read and follow this guide. It will produce excellent results.
http://www.avsforum.com/avs-vb/showt...7#post16166537

Thanks. Seems from Dan and other folks experiences the Epson's CMS should not pose the problems I've seen. CalMAN shows the Epson's colors are definitely over-saturated. I'm just having a bit of difficulty knowing how each setting, x/y/Y for each color is altered when I change the projector's settings.

Guess I'll have to keep playing.

Under RGBCYM the 3 sliders are interrelated. Hue is the least affected by the other 2, but saturation and brightness have quite an influence over each other. Any large changes in one control will almost certainly require an adjustment in the other.

Thanks. Now I'm wondering if I should initially aim to get the brightness for each primary and secondary right before working on saturation and hue. Or should I work each color separately for all 3 variables.

What would work best for the Epson?

I can't find the discussion here on using 75% saturation vs 100%. This was for an Epson and very informative.

Also, How do you get a 75% triangle in the CIE diagram?


thx

bob

From Tom's first post:

"Test Patterns
I provided a download link at the end of this post for a file from which you can create a calibration DVD with all of the necessary test patterns discussed here. There are two important rules to keep in mind.
1. When you use color and white test patterns, ensure that they are the same level of stimulus. For example, use 75% white test patterns with 75% color patterns and 100% white test patterns with 100% color test patterns. The same rule applies to windows and full fields. Use one or the other. Don't mix and match.
2. Use window test patterns only for CRT and plasma. For everything else, you can use either one."

Is that what you were looking for?

(Note: I'm only the most casual of C-HCFR users.)

As far as I know it's from/in Dan's spreadsheet.