Color HCFR Calibration Discussion (Post your calibration files here) - Page 127

Any guidance would be appreciated.

I've never calibrated a plasma, but your plots look very similar to my first front projector calibration attempts. In the case of projectors, color drop off at the high end is corrected by reducing the contrast setting.

Try turning down your contrast control a few notches, then recalibrate the greyscale, and see if the dropoff is reduced. Gamma should also improve. Keep reducing contrast a few notches and recalibrating greyscale until you gamma is flat at 2.2 and there's no color dropoff at the high end of the greyscale.

Hope this helps.

Steve,

I tried your suggestion and it did help greatly on my DLP and somewhat on my plasma.

Thanks!

Is it possible to use a correction matrix with another colorimeters beside the HCFR?
Or soemthing similar.

I can get a i1 Pro for a week end and would like to compare it with my spider 3 and make an correction mattic. But I can't select that feature with my spider 3.

Any suggestions what to do?

I have an i1 Display (predated the i1Display 2) and would like to get a calibration file usable in HCFR software.

I've tried a correction files suggested earlier in this thread and it got me a lot closer. Mine had a definate red push.

Does anyone know of a calibration service which will create a correction file that'll get me much closer? I've read about the guy in Canada but not sure his file will work on the HCFR. He appears to be creating files compatible with chromapure .

Thanks in advance.

duplicate post

duplicate post!

Yes, you can build a calibration file. Under Advanced --> Calibration File.. I think you can use 2 meter and profile one against another.

Usually, calibration file is target for specific device and environment. So, you are at the mercy of how close your environment/device is compared to the other that created for you. Also, each i1D2 behave differently, so your "correction" might require more/less.. At the end, it is a guess-game.
The best is of course for you to get a reference meter and profile against your own display. There are i1D pro rental program, but you are also at the mercy of how the person store/handle the device.
At the end, depend on how picky you are..

hi all, i've spent a couple of late nights calibrating my 43d450 (low-end samsung plasma) and i'm getting there slowly. my main problem is my greens are neon! to get to the stage below in my images i've increased green gain and green offset quite alot. i read that you shouldn't really touch them, but without increasing them i couldn't get some of the reds and blues low enough when calibrating the greyscale. as soon as i adjusted the greens i got much closer to 100% for r,g & b. edit: you can click on the thumbnails for a bigger picture.

what should i do differently? if i drop the greens back to standard, my red and blue won't get near 100%.

current settings:
13 - cell light
87 - contrast
57 - brightness
54 - colour
50/50 - tint

30 - red offset
39 - green offset
20 - blue offset
08 - red gain
42 - green gain
08 - blue gain

warm 1
gamma 0

everything else off

(for reasons unknown i didn't bring the hcfr file to work, but a load of screen shots instead- first lesson learned.)




this is the original screenshot on standard samsung movie mode:


this is the calibrated screenshot (camera moved a bit so not as sharp, but you can see the lovely glowing shrubbery)


edit 20/08: tonights calibration... colour dropped from 54 to 51, tint changed to 52/48

The luminance of your primaries is completely out of control.

red 23, green 88, blue 11, yellow 104, cyan 95, magenta 35, ...

Have you disabled all dynamic modes?

yes everything has been turned off. what figure should i be aiming for?

should i try and calibrate without touching the green controls and go for a best efforts?

You might want to start here:
http://www.avsforum.com/avs-vb/showp...postcount=1130
and follow the links as well.

The spreadsheet is quite handy and if you don't have Excel use the OpenOffice version, it is free.

thanks ill take a look. i should add that you can click on the thumbnails in my post for a bigger pic

When showing primaries/secondaries, click on the "down arrow" next to the "Editable data". That will show you "delta Luma" and you can see how much off you are.

Your luminance is off. Use this table for your reference. For Y, it is the % of White. In your case, If White Y=113.426, then your Green Y should be 0.7152*113.426 = 81.12 (you have about 87)
Blue Y should be 0.0722*113.426 = 8.19. (you have about 10.7)
Red Y should be 0.2126*113.426 = 24.11 (you have about 23.226)
So, you are about 10% too bright in Green, and about 20% too dim in Blue (hence the large DeltaE).

For primaries, get the luminance to be as close to each other...
If you have CMS, then try it. If not, you might need to reduce overall color down... it might affect your other color... but most device is a compromise. You should not expect you can get everything perfect, especially with a low-end device.
Other alternative is to get a video processor like lumigan or DUO or AVFoundry...

Rec. 709
x y Y
R 0.6400 0.3300 0.2126
G 0.3000 0.6000 0.7152
B 0.1500 0.0600 0.0722
Y 0.4193 0.5053 0.9278
C 0.2246 0.3287 0.7874
M 0.3209 0.1542 0.2848
W 0.3127 0.3290 1.0000

is the green too far off on the white control?

Remember that the green gain and green offset are adjusting the gray scale. That isn't the same as the color. Your gray scale looks very good. The luminance of green seems a bit high, which is what is likely making everything that is green look too bright. Unless you have a color management system, you won't be able to do much other than reduce the color control (basic color). The other thing to check is whether or not color is being decoded properly. I ran into this with one of my displays taking rec709 (HD) source material and misconverting it. The display applied a color 'twist' which made green much brighter. What is your source and how are you feeding it to the display?

ps3 via hdmi
hd burned to dvd
avs709

as angryht indicated, your grayscale is fine. No need to mess around with the RGB gain/offset anymore.

CMS is completely different. Your monitor might not even allow you to do this. YOu might need to search through the service menu or this forum to find out if your monitor have CMS or not.

Remember, the RGB gain/offset is to set the different RGB balance within the WHITE level.. Your Green looks too "bright"/neon because the individual color of GREEN is too bright. Most set do not include CMS.. you might try different color control (wide, standard, normal or something??) and that might help.

i've finished going over the colours and now the grass glows even more! i've edited post #3790 with the latest screenshot.

edit: i don't have CMS in the user menu (i've yet to check service menu) so adjustments done with colour and tint controls.

primary and secondary luminance:
red = 21.7 was 23 (target 24.4)
green = 81.7 was 88 (target 82.2)
blue = 9.6 was 11 (target 8.3)
yellow = 103.3 was 104 (target 106.7)
cyan = 90.5 was 95 (target 90.5)
magenta = 32.1 was 35 (target 32.7)

i concentrated on getting the green accurate at the expense of red (i tend to find images too warm anyway).

i even improved the greyscale before hand and gamma is now at 2.27 which was a nice surprise!

i've attached my hcfr file...

any more suggestions on how i can fix the neon grass?
i noticed that if i changed the image from WARM1 to WARM2 i can get much more accurate colours but at the expense of the greyscale.

 

Color_new_after.zip 4.4521484375k . file

thanks, that spreadsheet helped

If all you have are color and tint controls you're limited on what you can do.

Tint typically affects the location of the secondaries. If you subjectively think aiming for the 709 secondaries made the picture worse, then instead try to get the dotted lines on the CIE diagram to go from the primary, through white to the secondary. It looks like the dotted lines on the CIE diagram from the tint setting on the prior image were closer to white than on the latest chc file.

Color usually affects Y in comparison to white the most. If you're looking mainly at green, the delta luma line is basically dead on for Rec 709 on the chc file. The spreadsheet is under 1% for difference in Y for a custom color reference, so dropping color one setting may or may not be closer.

Note:
- Your measurements and evaluation video generally need to played from the same source. If you are playing the calibration video from a Blu-ray player, then you also need to use the Blu-ray player for evaluating other video. The reason for this stipulation is that the source can affect the video output, so if one of the sources modifies the video it will likely affect the resulting picture from another source. While ideally all digital sources should simply use matching video levels, that may or may not be the case, and you really don't know without further testing.
- When evaluating video it's a decent idea to use a few different videos, since it's unlikely to know how exactly grass was encoded with a single video.

Gamma measurements don't necessarily have much correlation with an actual on-screen image. If they did your measurements indicate that 60% would be darker than 50%, but if you look at a set of gray bars that probably isn't actually how your TV works. Anyway, it looks like the TV probably varies light output depending on the average picture level of the displayed image. You may get different Y readings by measuring a pattern that holds average picture level constant, and gamma is just a representation for Y values. This sort of variation may also affect Y values for color.

kpm197,

I hate to jump in here as I'm just now getting into display calibration.

One thing that may be a factor is if your colormeter has drifted. Mine has towards cyan. I can get the prettiest grayscale graph, yet the picture is fairly pink.

One thing that I see as a needed feature in display calibration is the ability to recalibrate the colormeter prior to use. I had a photography studio with an in house color lab. To check your process, you'd run control strips and measure them using an XRite reflection/transmission densitometer. Part of the routine would be to recalibrate the desitometer each time using an internal function that turned on a light of a know quality and having it reset itself to match that light output. Seems to me that there needs to be something similar in the display calibration field.

I'm aware that you can send some meters to have a calibration file created for certain color calibration software but what I'm speaking to is to have the self calibration as part of the device.

Attached is a list of four colorspaces showing the primaries and secondaries normalized to white (Y)
Rec 709 is the current HDTV standard.

Could not find the original and scanned from a printout and the OCR quit on me so attached the jpg instead.

Here is an old post, an early calibration of the LG 55LH90
http://www.avsforum.com/avs-vb/showp...91&postcount=4

Which is colorspace 1 and 2?

bob

IIRC (SMPTE-C) is the REC 604 SD standard, EBU is the European Broadcasting Union and read about DCI here:
http://en.wikipedia.org/wiki/Digital_Cinema_Initiatives

Question on all that. I get the colors x,y. But its Y that I ask about.

For Y white is always 1.0. With the other colors R being .21, G being .71, and B being .08

Now does this mean that somehow I have to get my colors to be exactly 1, .71 and .08 or does this mean white on my set is lets say 90 in HCFR so red should be 18.9 and G should be 63.9, B should be 7.2

I dont see how I would ever get the W, R, G, B to produce 1, and low decimal readings.

The tables are normalized to white (Y=1) so as to be independent of which calibration software and measurement unit fl or cdm.
Hence if white measures 30 fl multiply the color values by 30.
The spreadsheet linked earlier should do that for you.

If my advice is wrong no doubt some expert will chime in.

Those percentages are only correct if your primaries have perfect chromaticity (x,y). If not (99% of displays on the market) you need to compensate for any over/underage of saturation (and hue.) (Not to mention colorimeter crosstalk.) Minimizing dLuma will not always get to you the right answer.

In short, you can't correct just the polar origin point (D65 White) and not correct the remaining 99.9999% of color pallet. IOW, there is a reason why color filters have been used to set color/tint for decades ... think about it. If it turns out that your D65 White point actually contains 24% Red, 69% Green, 7% Blue then that's what you target, otherwise you no longer have a linear transformation.

Colorimeters are useful ... they can also be misused a/o misinterpreted.

The safest method is to use filters for all three colors. Get color/tint setting for Blue, Red, Green ... they will no doubt be different for each color. Once you have the settings for each color, simply take the average (mean) color setting and tint setting. This will get you to the optimum color/tint setting way faster than flailing around with a colorimeter. Then, if you have a working CMS and colorimeter/spectro, you can proceed to tinker further.

Note 1: You need a more robust color decoder pattern like the one from the AVS HD709 disk to get the best results from this method. The standard SMPTE style color-bar pattern is not the best choice here. Oh, normal color vision helps too.

Note 2: If you have to make a choice, it's better to optimize the Green-Yellow-Red side of the gamut triangle.

Edit: I suspect that I may not have explained things clearly enough ... go easy and try to clarify any misunderstandings ... engaging flame mode will get us nowhere.