or Connect
AVS › AVS Forum › Display Devices › Display Calibration › Basic Guide to Color Calibration using a CMS (updated and enhanced)
New Posts  All Forums:Forum Nav:

Basic Guide to Color Calibration using a CMS (updated and enhanced) - Page 9

post #241 of 1936
Quote:
Originally Posted by angryht View Post

I am pretty sure that HCFR measures Y candela per square meter (Y=cd/m^2). Doesn't 1 candela / (meter^2) = 1 lux?

no, lux=lumen/m^2 and is a unit of illuminance. nit=candela/m^2 = lumen/m^2/steradian is a unit of luminance. HCFR reports whatever the probe generates so if it's calibrated in lux, it will report lux, nits it will report nits. AFAIK it does not support any illuminance type meters.
post #242 of 1936
Thanks for the clarification, zoyd.
Quote:


HCFR reports whatever the probe generates so if it's calibrated in lux, it will report lux, nits it will report nits.

Maybe a dumb question but.....How do I know what units (nits or lux) my eyeone LT is generating? I though it was nits.
post #243 of 1936
Quote:
Originally Posted by angryht View Post

Thanks for the clarification, zoyd.

Maybe a dumb question but.....How do I know what units (nits or lux) my eyeone LT is generating? I though it was nits.

the default mode for the probes people are most familar (spyder, display LT, dtp-94, i1pro) is nits. Some have (I think the display LT is like this) an illuminance mode in which you place the probe into an opaque fixture and a different calibration table is loaded which will report in lux.
post #244 of 1936
That's what I thought. Thanks again, zoyd.
post #245 of 1936
Hello you experts. This is probably a dumb question : -) When you use a for example Pantone Eye-One Display LT to calibrate a projector where do you point it? Towards the projector? Towards the screen? If it is towards the screen will it not be a shadow from the measurement probe? Do you angle it so it does not measure its own shadow?
post #246 of 1936
Thanks so much for making this post. I've been searching for a primer on calibration and this was exactly what I needed. Much appreciated.
post #247 of 1936
Quote:


When you use a for example Pantone Eye-One Display LT to calibrate a projector where do you point it?

I pointed it at the screen because I saw others had done the same. If you point it at the PJ I believe you need a light diffuser and I think the one that comes with the LT is not good for that application. Angle it up (assuming the PJ is ceiling mounted) and towards the center.
post #248 of 1936
Quote:


Color Definitions (HD)
-------------x--------y-------Y
White--------0.313---0.329---1.00
Red----------0.640---0.330---0.213
Green--------0.300---0.600---0.715
Blue----------0.150---0.060---0.072
Cyan---------0.225---0.329---0.787
Magenta------0.321---0.154---0.285
Yellow--------0.419---0.505---0.928

If I'm calibrating a 1080p HD TV (DLP), but using a SD DVD player, should I use these numbers or alternative numbers for SD? Does it make a huge difference if I'm calibrating my HD TV using a SD source versus an HD source (I'll be buying a HD DVD / BR player shortly).
post #249 of 1936
Thread Starter 
Use these numbers. HD sources will be perfect. SD sources will be slightly oversaturated.
post #250 of 1936
Thread Starter 
I just finished putting together a spreadsheet that will automatically calculate Delta-E (CIELAB). All you have to do is input the xyY values. I thought that maybe some might find it useful.

 

delta-eCIELAB.zip 7.8486328125k . file
post #251 of 1936
Quote:
Originally Posted by TomHuffman View Post

Use these numbers. HD sources will be perfect. SD sources will be slightly oversaturated.

Thanks.
post #252 of 1936
i calibrated saturation and bias of primaries and seconaries and greyscale on my TW1000. The RGB gain and bias go pretty easy and i get them neatly on 100% from 30 to 100IRE(i sacrifaced somehow the lowest end becose the Spyder is known not reading it corectly)but get the color to the CIE triangle is pain in the a.. . I simply cant get the Blue right, even at maximum level it is still desaturated on the CIE chart(diagram attached). I use HCFR software PAL/SECAM set as reference(i live in Pal contry and overhelming majority of my dvds are Pal), HCFR pal disk and Spyder 2.
LL
post #253 of 1936
Quote:
Originally Posted by talkron View Post

i calibrated saturation and bias of primaries and seconaries and greyscale on my TW1000. The RGB gain and bias go pretty easy and i get them neatly on 100% from 30 to 100IRE(i sacrifaced somehow the lowest end becose the Spyder is known not reading it corectly)but get the color to the CIE triangle is pain in the a.. . I simply cant get the Blue right, even at maximum level it is still desaturated on the CIE chart(diagram attached). I use HCFR software PAL/SECAM set as reference(i live in Pal contry and overhelming majority of my dvds are Pal), HCFR pal disk and Spyder 2.

I wouldn't drive yourself crazy that still looks like a pretty good result. Be sure to check your decoder levels after moving all of those primaries on that machine. I had a TW700 once and after moving the primaries to the proper locations the decoder was royally screwed.

moe
post #254 of 1936
Zipplemeyer:you mean the Y values of colors and white? But what i can do, when they dont match the reference numbers?
post #255 of 1936
Thread Starter 
Quote:
Originally Posted by talkron View Post

Zipplemeyer:you mean the Y values of colors and white? But what i can do, when they dont match the reference numbers?

On a TW1000? Nothing, except use the main Color control to get one of them right. See the bottom of the original post.
post #256 of 1936
the control on TW1000 caled saturation? I suspected it has some other impact on the colors as saturation, becose changing it, does have no visible impact on saturation readings of the primaries and secundaries. My denon dvd player has some very simple color control too, turned up it makes picture cooler and down warmer...how it affects Y i can not tell...
post #257 of 1936
I know this is a little off topic, but I have a i1 display 2 (Gretatag) that I have used a couple of times tuning in my PJ. It is only a couple of months old and when I use the HCFR program, I get prompted to calibrate the colormeter on a planar surface, if I don't then the HCFR doesn't recognise the i1. I have not found what the planar surface it requires is, the best I can figure is a flat surface (0 light)? am I correct? I have found a lot of good information on this thread and hope you guys can help. Again sorry for veering off topic.
post #258 of 1936
Quote:
Originally Posted by Jason 1973 View Post

I have not found what the planar surface it requires is, the best I can figure is a flat surface (0 light)? am I correct?

yes, just block all light from entering probe.
post #259 of 1936
Quote:
Originally Posted by zoyd View Post

What I meant by that was when calibrating color you need to measure how far Y (as a percentage of white) deviates from reference to know if you have accurate colors or not. Color calibration has 3 dimensions to it (not 2) and the 3rd dimension (color lightness) has a large effect on perceived saturation. In some cases you can compensate by shifting the points slightly off of the diagram. The issue is described nicely at Bruce Lindbloom's site here.



So you need to measure Ycolor as a percentage of Ywhite and compare to reference, this may be why your colors still appear a litte "strange" to you.

Thank you, zoyd and Bruce Lindbloom, for this graphic depliction of the gamut. This post from a different thread made me curious about something. Because the color gamut is essentially a 3-D surface, if my primaries are undersaturated (inside the 2-D triangle), do the percentages in the beginning of this thread still apply. In other words, as this illustration shows the Z-axis, which I am considering is the Y (luminance) would be in a slightly different location if the x and y of the primary is not exactly on the reference (target) location. Is there an equation for the surface would vary slightly as the primary moves slightly towards the center?
post #260 of 1936
Quote:
Originally Posted by angryht View Post

Because the color gamut is essentially a 3-D surface, if my primaries are undersaturated (inside the 2-D triangle), do the percentages in the beginning of this thread still apply.

The percentages still apply since they are a reference standard. However, if your x,y location is off there are situations where you can trade-off errors in lightness for errors in saturation to achieve a net benefit in deltaE. You just need to use a deltaE formulation that includes both chromiticity and color intensity. Tom has a spreadsheet using the CIELab formulation above and Krasmusik has one for CIELu*v* in one of the sticky threads.
post #261 of 1936
Quote:
Originally Posted by zoyd View Post

You just need to use a deltaE formulation that includes both chromiticity and color intensity. Tom has a spreadsheet using the CIELab formulation above and Krasmusik has one for CIELu*v* in one of the sticky threads.

I have seen these spreadsheets. Is one prefered over the other, or should I try both. Do I just try to get the smallest deltaE?
post #262 of 1936
Quote:
Originally Posted by angryht View Post

I have seen these spreadsheets. Is one prefered over the other, or should I try both. Do I just try to get the smallest deltaE?

I don't have any empirical evidence to recommend one over the other although Krasmusik recommends Lu*v* as more appropriate for emissive technologies whereas Lab is more important for reflective (print) technologies. My guess is that at the level of accuracy of the standard probes either will do the job.

yes, minimize deltaE to get as close to reference standards as your display allows. btw, HCFR 2.0 will calculate the Lu*v* version so we won't have to use the spreadsheets in the (hopefully) near future.
post #263 of 1936
Quote:
Originally Posted by zoyd View Post

I don't have any empirical evidence to recommend one over the other although Krasmusik recommends Lu*v* as more appropriate for emissive technologies whereas Lab is more important for reflective (print) technologies. My guess is that at the level of accuracy of the standard probes either will do the job..

Alright, I'm going to have to slow you down a little bit. By emissive, do you mean things such as plasma, LCD... and other displays that emit light from behind screen. Would reflective (print) mean things like the screen from a FP, or a photograph?

Quote:


yes, minimize deltaE to get as close to reference standards as your display allows.

I will probably pick one and try to minimize the deltaE.
Quote:


btw, HCFR 2.0 will calculate the Lu*v* version so we won't have to use the spreadsheets in the (hopefully) near future.

Man, I can't wait for that!
post #264 of 1936
all displays are emissive - they emit lite. Reflective means the surface absorbs some light - reflects the rest - but they never emit lite. It becames debatable when you are talking about measuring a reflective screen under strong emissive lite rather than weak ambient lite if LAB or LUV is better suited. Sony has a prototype ebook that is the first of its kind reflective display - electronic ink on electronic paper - generates no lite.

gregr was of the opinion that if you primaries are off - then you would want to calculate what the luminance should be rather than using the REC709 standard.

But I think going with the dE is the better compromise - the main difference with LAB vs LUV in practice is the degree of error - LUV is better at gross error and not the imperceivable error - while LAB is better at the imperceivable error. It is absolutely crucial that McDonalds red be consistent in print media - they would lose their trademark if it was the wrong color - even if you cannot see it is different. That is what graphics uses LAB for. Whereas knowing how badly your TV is off - that is what LUV is for. You can use either one to derive LCH. For minimizing dE it does not matter - since in both systems the goal is 0dE. LUV and LAB have different dE measures though - so you should always specify what you used.
post #265 of 1936
Quote:
Originally Posted by zoyd View Post

The percentages still apply since they are a reference standard. However, if your x,y location is off there are situations where you can trade-off errors in lightness for errors in saturation to achieve a net benefit in deltaE. You just need to use a deltaE formulation that includes both chromiticity and color intensity. Tom has a spreadsheet using the CIELab formulation above and Krasmusik has one for CIELu*v* in one of the sticky threads.

If your primary locations are significantly off, then using the Rec 709 percentages to set your color decoder controls can introduce more error, not less. Greg Rogers posted an applet that will calculate the mix given a set of primary locations, so use that if you want to go the "match %Y" route.

If you are using a CMS, you absolutely need to use dE and be sure that the dE target is the reference primary you are trying to hit. For dE formula, you want to go with the 1994 version over the 1976. The 2000 formula is usable, as well, but interpreting what to do with the interaction term is a bit of a bear (of course, that's part of the issue folks are facing).


Bill
post #266 of 1936
Quote:


Greg Rogers posted an applet that will calculate the mix given a set of primary locations, so use that if you want to go the "match %Y" route.

Do you have a link?
post #267 of 1936
Thread Starter 
Quote:
Originally Posted by angryht View Post

Do you have a link?

http://www.accupel.com/

Click Manuals, and then you'll see a link for Luminance Calculator Programs.
post #268 of 1936
Thank you.
post #269 of 1936
That's a very handy program. Thanks very much.
post #270 of 1936
Thread Starter 
Quote:
Originally Posted by Bear5k View Post
If you are using a CMS, you absolutely need to use dE and be sure that the dE target is the reference primary you are trying to hit. For dE formula, you want to go with the 1994 version over the 1976.
Here's the CIELAB dE spreadsheet, except with the dE 1994 standard (the previous used the 1976 standard).

Bill, what is the advantage of 1994 over 1976? The only differences I see is that
  • the math is a lot more tedious to work out
  • the reported dE is much smaller
  • blue (and to a lesser extent green) accounts for a smaller % of the overall dE
For example, with the JVC RS1 (my current champ for chromaticity errors) using the 1976 calculation dE is
R: 25.4
G: 58.6
B: 33.1

and for the 1994 calculation dE is
R: 5.5
G: 12.7
B: 5.1

Interestingly, the respective gray scale dEs, where luminance is not part of the equation, are virtually the same for both standards.

 

delta-eCIELAB_1994.zip 12.4482421875k . file
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › Basic Guide to Color Calibration using a CMS (updated and enhanced)