Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: Broyer on March 08, 2007, 10:38:52 am
-
Hello everyone,
I am losing sleep over monitor calibration. Here's the deal:
HP dv9000t laptop, XP SP2. I calibrate with Huey Pantone. I running the latest driver for Huey and my video card.
My problem is this: When looking at Windows display properties, it lists the color management profile as the Huey:
(http://farm1.static.flickr.com/130/414648925_add5327b49_o.jpg)
This looks good to me.
But, when I go into the video card settings, Nvidia GeForce Go 7600 and select color correction, it lists it as STANDARD MODE. When I try to import the huey profile, I get a windows run32dll error and the whole window closes up. I can however import another profile, such as Adobe RGB 1998 etc... (but it looks like crap).
So, am I actually using the profile or not? My photos look great on my laptop (haven't printed any yet) but look lousy on other monitors.
(http://farm1.static.flickr.com/170/414648923_954bda8af0_o.jpg)
Here are two more photos for consideration:
a photo of my daughter shot with strobes. Converted to jpeg from raw in Bibble, CS2 to reduce image size, slight curve, slight sharpen.
skin tones look fabuluous on my laptop.
(http://farm1.static.flickr.com/87/414654086_b18102d2cc_o.jpg)
last photo is my colorchecker for reference. Nothing done to this except processed to jpg and reduced image size. No sharpening.
(http://farm1.static.flickr.com/162/414661468_65efe0b7db_o.jpg)
I'm lost. How does it look to you?
Am I actually using the profile?
Help please.
Tim
-
My problem is this: When looking at Windows display properties, it lists the color management profile as the Huey:
(http://farm1.static.flickr.com/130/414648925_add5327b49_o.jpg)
This looks good to me.
[a href=\"index.php?act=findpost&pid=105472\"][{POST_SNAPBACK}][/a]
At this time you are running the profile from the Huey, however, if you go into your video card and alter things there, and try to import the Huey into that, as you say it won't go.
The Huey is working after the video card but before the monitor itself.
So yes, at the above you are running the Huey and are calibrated to that
-
All you need to do is build a profile. It is automatically loaded at the system level. There is no reason to touch the video card settings. They are redundant and will only get you in to greater trouble.
-
All you need to do is build a profile. It is automatically loaded at the system level. There is no reason to touch the video card settings. They are redundant and will only get you in to greater trouble.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=105545\")
Jack, the OP has a point, there is no quick way to see how good a calibration is, without a reference. Luckily the OP has a reference. BTW, to random readers, Jacks's company markets some excellent calibration software used and loved by color geeks. In fact it's so good that I use it, after his co-worker spent an hour on the phone telling me what I should have read in the manual He should be *applauded* for helping to support one of his competitor's products.
Broyer,
here is a good way to test your monitor calibration, as you already own a Colorchecker which can act as a hard reference:
DOWNLOAD the following file.
[a href=\"http://www.babelcolor.com/download/ColorChecker_Lab_from_Avg.tif]http://www.babelcolor.com/download/ColorCh...ab_from_Avg.tif[/url]
Then open it in Photoshop (NOTyour web browser, that's another can of worms). Compare the Colorchecker to your screen. Happy ?
Edmund
-
Stay the heck away from the video card drivers. You will only undue what you have done.
Notice in Color Management it says "Default Monitor Profile: huey Unnamed Monitor?" That means it is using the profile you made and things are good. If you change the video card drivers that profile will mean squat.
If you want to test the profile, download the pdf "December 2004:Testing your display profile" from Andrew Rodney's site (http://digitaldog.net/tips/). Don't follow Edmund's advice as that will lead you nowhere.
You bought the heuy so you would not rely on your inconstant and inaccurate eyes for calibration. Why would you rely on them to verify a profile? The CC and the digital refference chart are two different mediums (reflective light vs transmitted light) and they will never match exactly. Even if you have a proofing setup.
Read Andrew's article. It takes less time, doesn't introduce dozen's of variables and it actually works.
-
Thanks All. I will try the advice out. Thanks for the digital dog link. There is a wealth of information on there. I'm reading the article now.
Tim
-
Compare the Colorchecker to your screen.
You've got to be kidding.
-
DOWNLOAD the following file.
http://www.babelcolor.com/download/ColorCh...ab_from_Avg.tif (http://www.babelcolor.com/download/ColorChecker_Lab_from_Avg.tif)
Then open it in Photoshop (NOTyour web browser, that's another can of worms). Compare the Colorchecker to your screen. Happy ?
Edmund
[a href=\"index.php?act=findpost&pid=105554\"][{POST_SNAPBACK}][/a]
Edmund, I can't believe you are suggesting this...
-
Edmund, I can't believe you are suggesting this...
[a href=\"index.php?act=findpost&pid=105654\"][{POST_SNAPBACK}][/a]
Go ahead, educate me
The file I referred to is Lab.
I never believed in the subtleties of monitor calibration testing anyway - I just take a Colorchecker, and a pair of female eyes, and mine own eyes, and look at the hard sample (Solux lamp or daylight), and the display sample, and compare. Yes, Elvis, there is such a thing as white adaptation.
Of course, I use a Colorchecker file in which I've painted in the *measured* values of my own Colorchecker chart rather than the generic average file referred to above. But I don't think the OP has a spectro.
Back when I was doing this more seriously, I used more complex tests using a spectro, but found these reflect reality less well -for me- than actual eyeballing as above. Amongst other problems with the measuring is the widely neglected fact that spectros and screens are polarized.
With two screens (Eizo, Samsung) on my desk, and a sample, a decent match among all three can be achieved if the white of one screen is measured and used to calibrate the other. Jack's Coloreyes software did this well already in the previous version.
By eyeballing you get a very good handle on the hues, *in a comparison*. Luminance gradation is a different game but maybe at this stage you have enough information to ridicule me already, and the OP has hung up out of boredom
Edmund
PS. Of course you can run various "evaluation" routines provided with the colorimeter software too; I am sure these are of interest to the makers of such software , and they trust their tests - I tend to trust my eyes first, in a *comparison*, and then to a more limited extent my spectros. Colorimeters are consumer products.
-
What videocard drivers refer to as "profiles" are some weird propriatary settings you can save in there. It's never used by anybody for anything. Just ignore them.
Your settings look good to me as they are.
If you feel like getting a closer look at XP color management you can download the color control panel applet. You really don't need it, but it can be fun if you're interested in the subject.
http://www.microsoft.com/downloads/details...&displaylang=en (http://www.microsoft.com/downloads/details.aspx?familyid=1E33DCA0-7721-43CA-9174-7F8D429FBB9E&displaylang=en)
-
Tim
Firsty stay AWAY from the NIVIDA drivers or you will end up in a double colour management siutation!
Secondly not sure why there are three profiles in the panel:-
AdobeRGB1998 - Why is this in here?? I though this was a colourspace rather than a colour profile.
Why do you have two huey profiles??
It looks to me like you may be loading several profiles one after the other. I only have the one colour profile in my settings (I am using a colorvision Spyder).
I would also check you startup folder to make sure you are not running the dreaded Adobe Gamma - this installs by default when you install Photoshop - get rid of it.
As for the advice of Edmund I'm afraid this seem like very odd advice to me - it may work for him but human eyes are very subjective which is why you buy a device in the first place.
Perhaps Edmund is just lucky that his eyes are accurate.
Andrew Rodney's articles are first class as are the late and missed Bruce Fraser. Also check out Digital Outback.
Good Luck
Jon
-
It looks to me like you may be loading several profiles one after the other. I only have the one colour profile in my settings (I am using a colorvision Spyder).
[a href=\"index.php?act=findpost&pid=106777\"][{POST_SNAPBACK}][/a]
There may be three in the list, but only the selected one is loaded.
-
That's correct. You can actually add more profiles from those stored in the Windows\system32\spool\drivers\color folder by clicking Add in the color management tab, but only one will be the default, as you select. When you color-calibrate with whatever product you use, it creates the profile, dumps it in this system folder, and sets it as default. When Windows boots, it automatically sets it up. If you run the calibrator multiple times and save with different target names, you can end up with various profiles in the color management list. In my dual-head system, both the calibrating app and Windows keep track of each monitor separately.
-
More information about CRT and LCD monitor calibration, ICC Profiles and Color Management can be found using these targeted swicki search engines:
Monitor Calibration Tips Search Engine (http://monitor-calibration-tips-swicki.eurekster.com/)
Color Management Search Engine (http://color-management-swicki.eurekster.com/)
-
I undertand you can have more than one profile in the list but for the sake of troubleshooting this problem I would get rid of all profiles bar the recent one from your calibration device.
Jon
-
I undertand you can have more than one profile in the list but for the sake of troubleshooting this problem I would get rid of all profiles bar the recent one from your calibration device.
Jon
[a href=\"index.php?act=findpost&pid=107015\"][{POST_SNAPBACK}][/a]
Windows will and can only load the profile selected as the default. For troubleshooting purposes we should concentrate on things that might cause trouble without suggesting false issues.
-
Go ahead, educate me
The file I referred to is Lab.
I never believed in the subtleties of monitor calibration testing anyway - I just take a Colorchecker, and a pair of female eyes, and mine own eyes, and look at the hard sample (Solux lamp or daylight), and the display sample, and compare.
I have no issues what so ever with this approach! Often, the best approach is simply KISS. This sure beats using the same instrument that calibrated the display to provide what is pretty much useless stat's such as deltaE compared to the target calibration aim points you asked for.
Using the same instrument to attempt to correlate some degree of accuracy, instead of a known reference instrument is kind of silly. Now using that instrument to provide a deltaE (which forumula?) over time, to gauge device stability IS useful assuming the instrument itself is a stable and repeatable device.
-
I have no issues what so ever with this approach! Often, the best approach is simply KISS. This sure beats using the same instrument that calibrated the display to provide what is pretty much useless stat's such as deltaE compared to the target calibration aim points you asked for.
Using the same instrument to attempt to correlate some degree of accuracy, instead of a known reference instrument is kind of silly. Now using that instrument to provide a deltaE (which forumula?) over time, to gauge device stability IS useful assuming the instrument itself is a stable and repeatable device.
[a href=\"index.php?act=findpost&pid=107212\"][{POST_SNAPBACK}][/a]
Fortunately we are living in a relative world and unfortunately you are applying absolute principles where they don't belong. No doubt Absolute accuracy would require some reference device of unquestioned accuracy. However using validations to test the monitor profiling function does not require that level of accuracy to provide valuable information about a profile and even more so about a series of profiles. The stats are far from useless. Clearly it can be seen that one profile can be more accurate than another based on a validation. It can be clear that higher validations might lead one to change their target values or ambient conditions, or a host of other steps in order to generate a lower set of delta e values regardless of the method or the use of the same instrument. To suggest that visually judging a color checker is a better method than a validation perhaps even sillier considering the instability of the human visual system. Clearly a reasonable instrument is more stable than that.
Now I suppose you could throw into doubt the quality and consistency of all the instruments on the market in which case we should all just go back to adobe gamma!
-
Fortunately we are living in a relative world and unfortunately you are applying absolute principles where they don't belong. No doubt Absolute accuracy would require some reference device of unquestioned accuracy. However using validations to test the monitor profiling function does not require that level of accuracy to provide valuable information about a profile and even more so about a series of profiles. The stats are far from useless. Clearly it can be seen that one profile can be more accurate than another based on a validation. It can be clear that higher validations might lead one to change their target values or ambient conditions, or a host of other steps in order to generate a lower set of delta e values regardless of the method or the use of the same instrument. To suggest that visually judging a color checker is a better method than a validation perhaps even sillier considering the instability of the human visual system. Clearly a reasonable instrument is more stable than that.
Now I suppose you could throw into doubt the quality and consistency of all the instruments on the market in which case we should all just go back to adobe gamma!
[a href=\"index.php?act=findpost&pid=107906\"][{POST_SNAPBACK}][/a]
You're again missing the point. We're not talking Absolute accuracy, that's a meaningless term.
Analogy. You use your foot to measure the patio you're building. I instead go to Home Depot and get a good old tape measure and compare your foot to a reference. Now the tape measure may be accurate to 1/100 of an inch but that's OK for the task at hand. I can now see your foot measures 11.3 inches. Your measurements for the patio are going to be an issue. I could buy a reference grade tape measure that's accurate to 1/1000 of an inch or even one that's accurate to 1/10000 of an inch. You might call the later absolute accuracy but I'm sure we have devices that can measure far finer units with far finer accuracy. There's a point of diminishing return here.
But if you use your left foot to gauge the accuracy of your right foot, you're simply fooling yourself. And this is essentially the functionality you're trying to sell your customers. Now, as I said, using the SAME instrument to measure something over time, assuming the instrument is repeatable to a published and useful degree (say less than a deltaE 2000 of 1) is useful for gauging device drift. Or, if you send the user a Minolta spectroradiometer to compare the data from your original device using the same algorithm (which may or may not have issues of their own), now we have something useful to compare. Has nothing to do with so called absolute accuracy as the Government or NASA probably has an instrument that's more accurate than the Minolta. But at least we're using differing instruments with specific degrees of accuracy to measure the process, not both your feet!
-
How about putting away deltaE numbers and use the good 'ole eyeballs to test color matching on a wide range of calibrated displays. I mean that's the whole point of color management and calibration. You know, one tagged device independant file looking the same on a wide range of devices as long as the devices are calibrated and profiled.
I finally came up with a more useable color target after a long troubleshooting session on this forum with a MacBook Pro display that turned out to be a display quality issue that I'ld thought had been remedied by Apple. It's a CM hue/saturation/brightness testing target using colors and hues greatly affected by display profiles built by software with bad chromatic transform formulas or basic corruption.
Give it a try and see if you get the same response on your calibrated display as I do on my EyeOneDisplay calibrated CRT which looks the same on my 2004 G5 iMac calibrated with Apple's eyeball calibrator containing a correct chromatic transform formula.
The instructions included I hope are clear enough. It requires 100% view size in a CM app-(no converting to sRGB for nonCM viewing), warmed up display and 6500K neutrality. It only has to be close to what's described, not perfect. However if the center squares look way too dark or light then there's something really off with your calibration and/or profile.
-
You're again missing the point. We're not talking Absolute accuracy, that's a meaningless term.
Somebody is missing the point, thats for sure. There is not a day that goes by that we don't successfully employ validations to troubleshoot customer calibration problems. They are an incredibly useful tool for testing various target values against one another as well as ambient conditions and a host of other conditions. You need to get off the sale guy baloney and look at the real world were these things are being done and providing valuable feedback. Never would I nor have I suggested that they can be used to judge accuracy within any stated percentage. So please don't sit there and tell me I can't get any value out of something we employ all the time to assist customers after THE SALE to improve there profiles and profiling habits. I get your point, it's just not relevant to the way the tool is being applied in this case.
-
Oh as to the stat's being useful for deciding what target values to use, again, I don't buy that.
All we're doing here is sending known color values, probably in LAB to a device then measuring those values using said instrument and comparing what we ask for with what we get. The process doesn't have any way to tell you that a native gamma and white point will provide less banding or that there's some correlation between the ambient light and the display plus how you view the prints are ideal although it can provide some suggestions based on some very old ISO specs. It has no idea what printer profile will be used nor if you've setup the soft proofing correctly to handle the dynamic range of the paper. All this validation can really do is compare the measured data with the data sent by the software but since you're using the same instrument, its again like using your left foot to gauge the accuracy of your right foot with no other reference grade measurement device to tell you both are off by X amount. X amount MAY be acceptable! But we simply don't know that.
Lastly, as Edmund mentioned, loading an image with a soft proof and examining a reference print is a much more effective way to see if all your ducks are in order here. We want the print and display to match as closely as possible based on their differences in reference media and dynamic range.
There are probably all kinds of way to measure the print versus the display but if they don't appear to match, what's the point?
Also, we're NOT really producing a white point that's D50 (or any other standard illuminant). If that were the case, everyone would have accepted years ago that calibrating a display to D50 and having an output profile that assumes D50 would match. Yet years of work have resulted in people calibrating to D65 (or god forbid, 6500K) while nearly every printer profile assumes a D50 viewing condition. And the only real D50 viewing condition comes from a light source 93 million miles from your light booth. So KISS does work and we shouldn't put so much credence in measured values all the time. They ARE useful, but they are also often not.
-
Yet you refuse to define accuracy or how using the same device to measure itself is accurate.
Define accuracy. Its one of those buzz words used to sell something to people who feel they need it but what does it mean? Accuracy based on what standard measured by what, itself?
If you want to tell people your left foot is 12 inches, so be it. How accurate is that without having something else to compare those measurements to? Its like the nonsense that a camera profile produces accurate color. That's ridiculous.
Accurate color is colorimetrcally correct which means the measured color. But measured by what and what's the accuracy of the measurments? That we CAN define by using other instruments of known quality. But we can't define it using the same instrument unless we just want to make ourselves feel good.
When you buy a car, there's a gas mileage associated wit the car. Its useful to compare a Hummer and a Prius. But there's always the fine print. Your mileage may vary!
-
How about putting away deltaE numbers and use the good 'ole eyeballs to test color matching on a wide range of calibrated displays. I mean that's the whole point of color management and calibration.
Exactly! But that's not very sexy.
There are issues where using your eyes can fool you but that's usually not when viewing images in context. And as you say, the point of all this is to provide a reasonable match between two very dissimilar media. It will never and simply can't be prefect. We've got glowing phosphors (or a Fluorescent backlit product) and a reflective print with usually quite dissimilar dynamic range. Ain't ever going to match 100% even when we have Star Trek or Star Wars technology. All the Bells and Whistles only complicate the process for many users. The geeks love it. But the bottom line is, do the two match?
-
Ok, lets use your analogy. You build a series of walls using your right foot to measure. At the end of the day you have three walls the same and two that are different. So did your foot change size or did you screw up the placement of your foot while measuring. This has nothing to do with your left foot at all. We have to accept that the instrument we buy is reasonably accurate and reasonably consistent. If they are not we should throw them out.
As for stats being used to determine the best result, you are making the suggestion that all displays behave exactly the same as every other display at one set of target points and that just ain't so. Luminance, gamma, white point and black point target values should all be fine tuned to the particular display, more so the cheaper they get. If you ignore the display's behavior and capabilities again we might as well use adobe gamma. I hear all the time about what the default standards are or should be and I watch customers struggle to hit target values they simply can't. Validations are indeed a way to make that point.
-
Ok, lets use your analogy. You build a series of walls using your right foot to measure. At the end of the day you have three walls the same and two that are different. [a href=\"index.php?act=findpost&pid=107923\"][{POST_SNAPBACK}][/a]
They are the same and I didn't say they wouldn't be. But are they the right measurement? I asked for a wall that's 12 feet, you gave me a wall that's 11.5 feet. I asked you to prove to me my wall is 12 feet. So you measured it again with your foot. I don't buy that as being useful. I instead use the tape measure from Home Depot and guess what, my wall isn't 12 feet.
Again you've missed the point and the analogy. You're software and your foot are supposed to tell me the accuracy of the measurements. But they don't.
-
Actually the single foot anology is correct for this situation. We have chosen a device to calibrate with. We rely on it to be reasonably accurate. I've chosen my foot as the standard. So long as all the walls match I'm using the tools correctly. You are asking for something else entirely which is proof that it is accurate within some standard. I'm counting on the manufacturer to provide that and am using that reasonable assumption to create some trending that does indeed tell me alot. You're asking for something you can't have and neither can any of the readers of this forum. Again back to my suggestion of absolute verses relative. I'm relatively comfortable that the instrumentation we have available is accurate enough. With that assumption I can build some trending that I find valuable. You want higher precision and you can't have it so you'd rather throw the baby out with the bathwater and suggest that only a visual test will do. I understand exactly what you are trying to say. I just don't agree plain and simple.
"You're software and your foot are supposed to tell me the accuracy of the measurements. But they don't."
This simply isn't true. You are being absolute again. There are degrees of accuracy and I'm saying we hit a high enough standard to be valuable and you are saying only the application of a third device will yield any relevant data. And I'm saying if that were really the case we should stop profiling and use adobe gamma because if you're right all these instruments should never be used by anyone to profile until each and every one has been tested on site.
-
Actually the single foot anology is correct for this situation. We have chosen a device to calibrate with. We rely on it to be reasonably accurate. I've chosen my foot as the standard. So long as all the walls match I'm using the tools correctly.
[a href=\"index.php?act=findpost&pid=107950\"][{POST_SNAPBACK}][/a]
OK so exactly what are you measuring and what are the results supposed to tell the users?
You send say 10 color patches to the display with 10 lab values. You measure them with the instrument. Then you measure them again after the calibration process. You compare the deltaE of the 10 patches with the same instrument. So you're gauging what accuracy? We expect that if you measure the 10 patches and get 10 values, then measure them again, you should get the same 10 lab values (within reason). Since you used the same instrument, we don't know how accurate either the first set of 2nd set is based on some better standard. So this begs the question, just what did I gain here?
-
I'm relatively comfortable that the instrumentation we have available is accurate enough. With that assumption I can build some trending that I find valuable.
[a href=\"index.php?act=findpost&pid=107950\"][{POST_SNAPBACK}][/a]
I said in my very first post that trending is useful. And I agree that most instruments are 'accurate' (within tolerance of what a human can perceive).
Trending is good, no question. What else are we supposed to gather from the stats?
-
Alrighty then...
Andrew, I agree with you in our point on not using the same device to calibrate and verify. It's a no-brainer. I also agree that the level of "accuracy" does not need to be great. However, I disagree with the idea that looking at a virtual Gretag Color Chart on the monitor and comparing it to an actual CC is effective means of gauging a monitor profile for the following primary reasons:
1. Not everyone is skillful at gauging color and/or do not understand how human perception can be effected. They may think things are fine when someone like you or I would see a noticeable difference.
2. Not everyone has a working area suitable for such a test. Lighting that varies over time or is a deranged mix of light sources, bright lime-green walls, brightly colored pictures or decorations, stained wood desks, clothing being worn, etc. I've seen it all and these things have an effect on how our eyes perceive things. A skilled operator could manage in such an environment over time but many people don't have that ability yet.
I think comparing a CC to a virtual CC on screen can be just as ineffective in gauging monitor calibration/profiling as using the same instrument for both profiling and verifying. While a skilled viewer may be able to do as you describe effectively, normal people (the average person) is not in that position. You and I and others like us are not normal people.
What about your article on monitor profile testing in photoshop? Wouldn't that be a more effective means of gauging if a profile may have issues or not?
-
I said in my very first post that trending is useful. And I agree that most instruments are 'accurate' (within tolerance of what a human can perceive).
Trending is good, no question. What else are we supposed to gather from the stats?
[a href=\"index.php?act=findpost&pid=107956\"][{POST_SNAPBACK}][/a]
Fuinny. Perhaps if you paid attention to what i was saying instead of what you thought I was saying you would see that trending is exactly what I have been saying all along.
and that , that trending can be used to interpret and modify choices to acheive a better profile. As I said it fits the one foot analogy.
-
1. Not everyone is skillful at gauging color and/or do not understand how human perception can be effected. They may think things are fine when someone like you or I would see a noticeable difference.
2. Not everyone has a working area suitable for such a test.[a href=\"index.php?act=findpost&pid=107958\"][{POST_SNAPBACK}][/a]
Agreed on both counts. This would most certainly be a very subjective evaluation but, the same person viewing both is probably going to be using the system so it has some merit.
There's nothing we can do about the percentage of color blind men and I don't suspect we'll be using the Munsell tests before such evaluations.
There are a number of tests one can use to help evaluate the calibration of a display (the old black screen with selection in the middle: Where do you see zero black separate, do the steps appear neutral). There's the full screen 100% zoom on a black to white gradient with display profile assigned to the file and so forth.
In the grand scheme of things, if the user feels the screen and print match to an acceptable degree, we're in pretty good shape, even if that person's wife can see they don't match ;-)
-
Fuinny. Perhaps if you paid attention to what i was saying instead of what you thought I was saying you would see that trending is exactly what I have been saying all along.
[a href=\"index.php?act=findpost&pid=107963\"][{POST_SNAPBACK}][/a]
Funny, if you read what I wrote, I said that in my first post. Then again lower down.
That doesn't explain this:
Clearly it can be seen that one profile can be more accurate than another based on a validation. It can be clear that higher validations might lead one to change their target values or ambient conditions, or a host of other steps in order to generate a lower set of delta e values regardless of the method or the use of the same instrument.
So what's this got to do with trending which is comparing the SAME target values over time? It tells me there's a difference between the condition of the device today compared to a week ago. Now how does that tell me I should change my target values?
How do you decide the profile is 'accurate'? We can see that the device has changed since the last calibration session. How does that correlate to profile accuarcy in the first place?
Or:
There is not a day that goes by that we don't successfully employ validations to troubleshoot customer calibration problems. They are an incredibly useful tool for testing various target values against one another as well as ambient conditions and a host of other conditions.
How does comparing the device drift have anything to do with the validation of the target values? It tells me they change, which is what we expect or we'd only calibrate a display once and be done. The validation is useful to tell you, you need to recalibrate OR you should calibrate every such and such number of hours. But validate the initial targets? How does this work?
If a customer is having an issue, they could run validation to save a few minutes in which they are informed they need to recalibrate. Or they could just recalibrate! Where's the bit about profile accuracy? The profile was presumably accurate once you finished the original calibration and profiling.
and that , that trending can be used to interpret and modify choices to acheive a better profile.
How does one interpret this to modify a better profile? Better how?
-
Boy, I've never seen a thread more aptly titled .
-
Why are we talking about device drift here? The variation in accuracy from profile to profile is far greater with different target parameters, ambient conditions and monitor quality. If devices drift that much we should all stop profiling altogether. Why are you suggesting all the drift is in the device? Monitors are so much more unstable compared to a dtp-94 or eye one two so we shouldn't even be considering them as a critical problem.
Based on all the things you say are true, everyone reading this list needs to stop profiling RIGHT NOW. If you can't verify at any level of accuracy at all then you probably can't build either so lets all just quit.
-
Why are we talking about device drift here? The variation in accuracy from profile to profile is far greater with different target parameters, ambient conditions and monitor quality. If devices drift that much we should all stop profiling altogether. Why are you suggesting all the drift is in the device? Monitors are so much more unstable compared to a dtp-94 or eye one two so we shouldn't even be considering them as a critical problem.
Based on all the things you say are true, everyone reading this list needs to stop profiling RIGHT NOW. If you can't verify at any level of accuracy at all then you probably can't build either so lets all just quit.
[a href=\"index.php?act=findpost&pid=108050\"][{POST_SNAPBACK}][/a]
Accuracy based on what? Accuracy based on what? How many times do I have to ask you.
The same instrument measures the behavior of a device over time just like you measured three walls with your foot. So the accuracy of the device isn't what we're looking it, it's the accuracy of either how well the device behaives in the SAME condition over time (device drift) or you'll tell us what you're measuring with the same instrument that provides a delta of accuracy of the profile.
I told you how I assume you're measuring reference colors to actual measured colors. What else can you do? So taking device accuracy out of the equation, just what on earth are you measuring for accuracy and how given we can't gauge the device itself like your foot.
And no, based on what I'm saying it doesn't mean people shouldn't calibrate their displays. We all know they drift and we have to put them back into a stateded condition based on measuring a pile of colors. But where's the accuracy in YOUR statement about profile accuracy and feedback on target values?
-
OK just as a reality check, I called Karl Lang, the color scientist who designed the Radius PressView and later the Sony Artisan about this so called “accuracy test”. His opinion was slightly different than mine. I said it’s mildly useful. He was more forceful (its useless) at least with respect to the 'accuracy of the profile'.
Lets talk true validation. We had this in the Artisan (Quick Calibration). A number of known color values are sent to the display after calibration and profile building. The instrument compares the values based on the measured data. The idea is to tell you if the device has altered its behavior to a fixed deltaE such that you should recalibrate. What more modern products have done is simply store this reality check and track this graph, telling you how far the device has deviated over time. This is useful in telling you that calibrating the device once a month isn’t frequent enough (from month 1 to month 2, your deltaE (say deltaE 2000) is 3, you then find that doing this process weekly provides results that are less than 1, its good indication you should do this more often).
In the case of the Artisan, it took 12 minutes for a full calibration. The Quick Calibration would take no more than 7 minutes. IF the deltaE was too high, it would instead run a full 12 minute calibration. Its a time savor and its useful to do before color critical work. This isn’t about accuracy because again, we’re using the same instrument, and software to measure a subset of colors. Otherwise, run the entire process and just build a new profile.
OK now onto ‘accuracy’. Lets see the definition:
The state of being accurate; freedom from mistakes, this
exemption arising from carefulness; exact conformity to
truth, or to a rule or model; precision; exactness; nicety;
correctness; as, the value of testimony depends on its
accuracy.
2: (mathematics) the number of significant figures given in a
number; "the atomic clock enabled scientists to measure
time with much greater accuracy"
In the case being discussed here, Jack (and to be fair, all other’s producing software to build profiles) often use the term profile accuracy. What’s it mean? To build a profile be it for a printer or a display or capture device, known color values are sent to or captured by the device. They are measured and a comparison of known and produced LAB values are provided. This allows one to build an ICC profile. In the case of a printer, one could send a known value to the output device based on the profile, measure it and compare the LAB values. But now you’re back to the issue of using the same device! The instrument has a fixed and specific illuminant that may be totally different from the illuminant under which the print is viewed. And heck, do you like the way the print appears in the lighting you’ve built the profile for based on how the image will be viewed? This goes back to the suggestion of just looking at images on the display and comparing them to the print. Do they match? Keep in mind that printer profiles are pretty complex. They have multiple tables for handling different rendering intents and they have to provide a soft proof as well. So there’s the output you get AND the values sent to the display profile for soft proofing. Makes discussing display calibration with respect to a soft proof a lot more variable and difficult. There are some tricks for examining the deltaE of printer profiles by comparing round trip errors going though the PCS. But ultimately you just send a lot of images through the profile, make prints and LOOK AT THEM. The Perceptual mapping is solely based on pleasing color. There is no fixed specifications for how a profile vendor can or should build a perceptual table. And try using an Absolute Colorimetric intent for output (which should in a prefect world produce absolute colorimetric accuracy) and you’ll see a print that’s pretty butt ugly.
The bit about profile accuracy for the display could be determined but NOT with the same instrument that built the profile as I’ve illustrated. If we send 50 solid patches to the display and measure them, how accurate are the resulting readings? Only when you use a known reference instrument that we KNOW has a higher level of accuracy (those significant figures given in numbers), can you know that the original 50 values are accurate and to what degree numerically.
So, how does measuring a small sample of patches, the case with all display profiling products, using the same instrument tell us the profile is accurate to the target values we’ve asked for? In a perfect world, we’d measure 16.7 million samples, one for each possible color. The profiles would be HUGE. It would take forever to measure. In the case of a printer, one can generally produce an acceptable profile using 900-4000 patches of colors. All the others are for lack of a better word, extrapolated to build the profile (which can define 16.7 million colors). For a display, far, far fewer patches are measured. So we have a lower number of samples to measure and we’re measuring it using the same device so there’s no way to measure the accuracy of the profile. We can measure the differences in each profile built over time to gauge device drift but that set of measurements may not be ‘accurate’ to a higher measured standard and that’s OK. As long as the device is consistent (and we assume they are), the inaccuracy over each group is fixed and what we’re trying to measure here is the difference over time, not the accuracy of the original or subsequent profiles.
Accuracy is a marketing buzz word. It’s used to sell stuff. And ALL the color management companies are guilty of doing this. This isn’t any more correct than years ago hearing color management companies sell their wares using the term ‘push button color’.
I’ve asked Jack a number of times how his process gauges accuracy based on the facts above. How does the instrument along with some sample of known and measured LAB values tell you how to set the target calibration (which on an LCD is limited to the intensity of the backlight). If the soft proof seems off, using validation CAN tell you that your current profile isn’t accurately describing the current behavior of the device. The device has changed so trash the profile and start again. But short of that, how can sending X number of LAB values tell you anything more? Where’s the accuracy? What’s the software supposed to be telling you? Still waiting on those answers.
Our job as consumers, (and educators) is to separate the facts from the fluff. To decide if functionality provide in a piece of software is useful or there as a feel good placebo ( an innocuous or inert medication; given as a pacifier or to the control group in experiments on the efficacy of a drug).
-
After reading this thread, I'm definitely ready to stick someone's foot SOMEWHERE. I think I got lost after the section on comparing deltaE to my nicotine patch. Now, can someone tell me, who's foot is BIGGER? That's the man for me.
I sure wish I had Karl Lang's phone number to call when I had a pissing contest to win. Thanks for the laughs, guys. Always good to know we count on you to teach us something. Now it's off to Home Depot for me. I hear they have a sale on monitor calibrators and I hear the guy at the paint counter comes with the most accurate set of eyes. Wish me luck!
Misty
-
I have duck shaped feet, mistybreeze. Sorry to disappoint.
Why all of a sudden do I feel so inadequate?
-
Ok lets apply a real world example to this silly discussion. This one happens every week like clock work. I get a valiudation screen from a customer and all the dark grays and blacks are delta e's of 4-20. My first question is what are your ambient conditions and everytime it's bright. Simple deduction, turn off the lights and run again. Lo and behold the values drop dramatically. Here's another one. Customer sends a validation screen where most of the values are way above 3. The target luminance he is trying to hit is 90 and he has a nice shiny new eizo ce. Change the target value to 150 or so and again the validations drop like a rock cause there is no way that monitor works at below 120 with any accuracy at all. Now neither of these is primamrily instrument drift and it is fair to say the the results after the change are far more accurate than before. Now without some sort of validation I can't imagine how you can confirm any problem condition like these.
Again lets all remember that I have used the words relative accuracy repeatedly in this conversation
So while you and Karl are no doubt real smart guys you are ignoring the practical value of a tool that consistently delivers valuable information.
-
Ok lets apply a real world example to this silly discussion. This one happens every week like clock work. I get a valiudation screen from a customer and all the dark grays and blacks are delta e's of 4-20. My first question is what are your ambient conditions and everytime it's bright. Simple deduction, turn off the lights and run again. Lo and behold the values drop dramatically.
And you had to tell them that, the instrument and software didn't? And what got measured to produce this detlaE of 4-20, the before then after ambient light? What process did the user conduct to get those values? He measured what? He held up the device the first time and got an ambient reading then a week later? I'll buy that but it's NO different from holding the device to the display and comparing the LAB values in both instances so it IS measuring device drift (the device is the lighting around the display). We've already said comparing original and updated measurements with the same device is useful and has nothing to do with profile accuracy.
Here's another one. Customer sends a validation screen where most of the values are way above 3. The target luminance he is trying to hit is 90 and he has a nice shiny new eizo ce. Change the target value to 150 or so and again the validations drop like a rock cause there is no way that monitor works at below 120 with any accuracy at all.
Same question as above. What was measured to produce these results? Why did the software not 'tell' the user via this magical deltaE in the first place it can't produce a luminance of 90 cd/m2? He asked for 90 and the software did what? Not warn the user the display can't get to that target?
Now neither of these is primamrily instrument drift and it is fair to say the the results after the change are far more accurate than before.
You sure love to use that word accurate. Accurate to what? Sounds like in one case the user had two vastly different ambient light conditions which the software didn't detect, the user saw a visual problem and you had to tell him this. Or the software should have told him at the get go. In the 2nd, the software and hardware didn't know enough to tell the user 'you can't get that target value'. In each case, I still don't know how you define what's accurate.
If someone calibrates their display and opens a soft proof, then I shine a 10,000 candela light in their eyes and they report a problem seeing the display, I can probably inform them they have a viewing condition issue (and maybe they could hold up the puck to tell them that too which I guess is mildly useful in stating the obvious). But where's the accuracy you keep using? And how does this tell the user in your examples how he should have set the target calibration aim points which you say it did? Again, what was measured initially and subsequently here that told the user the problem, not the guy in tech support?
Again lets all remember that I have used the words relative accuracy repeatedly in this conversation.
Relative to what?
So while you and Karl are no doubt real smart guys you are ignoring the practical value of a tool that consistently delivers valuable information.
[a href=\"index.php?act=findpost&pid=108414\"][{POST_SNAPBACK}][/a]
That's what everyone here is trying to decipher. Other than device drift based on the differences in a set of measured lab values over time, something we all agree is useful, just what information is being provided and how???????
And ALL hardware devices supported measure ambient light?
-
I continue to use the words "Relative accuracy" This discussion has become a huge waste of time because you want me to place some specific values on the table so you can try to hit them. As I have said over and over and over again we use verifications to compare one profile to another in a relative sense. Above I gave two simple illustrations of how you can use verifications to compare results of different target values as I have said at least twice and probably 20 times above. It really doesn't matter how accurate they are, within reason, if they are reasonably consistent. You can troddle on and on about them being absolutely worthless but as usual you will ignore practical experience over your perception of something you have not tried.
And as usual you did not read what I said. For instance I did not say the software/monitor could not produce a luminance of 90. I said when you hit a target of 90 the monitor performed badly. I did not say the customer measured the ambient light. I said he had high ambient light conditions when he built the profile.
Feel free to ramble on about my short comings. I'll be going back to helping our good customers get good profiles.
-
This discussion has become a huge waste of time...[a href=\"index.php?act=findpost&pid=108570\"][{POST_SNAPBACK}][/a]
Yet, you continue.
Here's my definition of turn-off: an intelligent man who permits his ego and stubbornness to diminish his ability to demonstrate just how smart and helpful he can be.
Some of you guys do a great disservice to yourselves and your erudite reputations by displaying such childish insecurity and pettiness. Ignoring this thread is like resisting the temptation to pull out a camera and photograph a car wreck when you normally shoot nothing but beauty.
Here's my definition of sexy: an intelligent man who's secure with himself, who doesn't have to win every debate or strike back at every challenge and/or criticism. Sometimes he comes with a cute face. Where's Karl Lang when you need him? No wonder he doesn't post here.
-
I continue to use the words "Relative accuracy"
Without any definition of what that means other than the term accuracy sounds like its providing an end user something that I'm trying to determine is useful or simply hype.
This discussion has become a huge waste of time because you want me to place some specific values on the table so you can try to hit them.
I can't hit anything that you refuse to explain or define. But I agree, this is a huge waste of time. You have been quoted a number of times in this one series of posts discussing profile accuracy and so on, without ever answering what the heck you're trying to define. This is Orwellian speak at its smoothest.
As I have said over and over and over again we use verifications to compare one profile to another in a relative sense.
So that has nothing to do with profile accuracy? It does as I've said from the first post allow us to define device drift (which you then suggested was incorrect, brought up Adobe Gamma and suggested based on my argument, we should not calibrate our displays.
Above I gave two simple illustrations of how you can use verifications to compare results of different target values as I have said at least twice and probably 20 times above.
But when called to explain the holes in the argument or clarification, you bypass the questions which is makes me more highly suspect of your thinking here.
It really doesn't matter how accurate they are, within reason, if they are reasonably consistent.
Accurate to what? We both agree the instrument is or should be repeatable. We both agree that taking two sets of measurements over time can provide a deltaE of device drift. Which of the two profiles is 'accurate' and to what? That's the question I've asked you since day one that you skirt.
I don't of course expect an answer this late in the game.
You can troddle on and on about them being absolutely worthless but as usual you will ignore practical experience over your perception of something you have not tried.
What would make you think I haven't tried it?
For instance I did not say the software/monitor could not produce a luminance of 90. I said when you hit a target of 90 the monitor performed badly.
And I asked why the software didn't tell the user at the get go it couldn't produce the luminance asked and how a validation process which compares a reading made earlier and one just made is helping here or how that has anything to do with profile accuracy. Why didn't the software pop a warning telling the user he/she can't produce that target from the first session? And what does 'preform badly' mean?
Feel free to ramble on about my short comings.
You're doing fine on your own.
Sorry Misty, I'm done. It appears that if you ask for simple technical explanations after a vendor posts about his product that sounds either unclear or suspicious, the posts go on and on and on in an attempt for clarification. I should let it all rest, assuming people here are smart enough to see thorough the hype. But after all these years of hearing so much junk from color management vendors, it ruffles my feathers.
-
Sometimes he comes with a cute face. Where's Karl Lang when you need him? [{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=108580\")
Well here's his cute face:
[a href=\"http://www.lumita.com/]http://www.lumita.com/[/url]
-
Well here's his cute face:
http://www.lumita.com/ (http://www.lumita.com/)
[a href=\"index.php?act=findpost&pid=108597\"][{POST_SNAPBACK}][/a]
Interesting that the cute face of this color guru is shown in B&W.
-
Andrew, Jack, let's stop doing this in front of the children
We've all stated our opinions here, we all actually have high respect for each other on a personal level, and it's time to move on.
I would like to thank Jack for unselfishly supporting users of rival products on this forum, when they are in distress, and congratulate him for his excellent product.
I would like to thank Andrew for helping people here with his broad expertise of color management, giving freely of his time in order to make things work better.
Thank you also to people who helped with Windows and driver-specific information for the original poster's concerns.
I do not feel so impressed by some other members of this forum who fanned the flames - you know who you are.
Edmund
-
Interesting that the cute face of this color guru is shown in B&W.
[a href=\"index.php?act=findpost&pid=108607\"][{POST_SNAPBACK}][/a]
Yes but at least its accurate <g>
-
Yes but at least its accurate <g>
[a href=\"index.php?act=findpost&pid=108661\"][{POST_SNAPBACK}][/a]
Or is it "relatively accurate?"
-
Andrew, Jack, let's stop doing this in front of the children
We've all stated our opinions here, we all actually have high respect for each other on a personal level, and it's time to move on.
I would like to thank Jack for unselfishly supporting users of rival products on this forum, when they are in distress, and congratulate him for his excellent product.
I would like to thank Andrew for helping people here with his broad expertise of color management, giving freely of his time in order to make things work better.
Thank you also to people who helped with Windows and driver-specific information for the original poster's concerns.
I do not feel so impressed by some other members of this forum who fanned the flames - you know who you are.
Edmund
[a href=\"index.php?act=findpost&pid=108639\"][{POST_SNAPBACK}][/a]
I've gotten past wondering why you seem to feel as though anyone cares about your non-opinions only to find myself wondering how you ever manage to speak at all given the number of dicks you're constantly struggling to shove in your mouth at once.
And is there any forum on which the ColorEyes "staff" have not made complete jackasses of themselves? Considering the fact that they post exclusively as a marketing tactic, you'd think they'd at least attempt to cover up their own ignorance as to how the software they didn't write and have very little to do with actually works. Given the quality of the code their company mysteriously churns out I'm beginning to suspect Edmund is their chief engineer.
And why haven't I been banned for my "personal attacks" against the idiots shitting up this entire forum yet? It'd be a lot quicker than deleting my posts since I clearly lack the self-control necessary to resist the urge to bitch when the S/N ratios drop to these dpreview levels. Am I not swearing enough or something?
-
And why haven't I been banned for my "personal attacks" against the idiots shitting up this entire forum yet?
[a href=\"index.php?act=findpost&pid=108745\"][{POST_SNAPBACK}][/a]
Just keep in mind not everyone with Irritable Bowel Syndrome is an idiot. I know a few "loose cannons" who are quite brilliant if you can catch them outside a men's room.
Orangekay, I sure hope nobody bans you. I love a good cuss sergeant. Your virulent posts are so well written and nobody defines "angry bitch" quite like you. After reading your posts, I feel downright angelic, as if I've just gone to confession and prayed the rosary as my penance.
Knock 'em dead, precious. Maybe you should re-register with a new name: I'm thinking BloodDiamond. They don't invite me to the marketing round-tables for nothing.
-
We sure like a good arguement, don't we! And to think, all this vitriol has resulted from a question about calibrating a laptop.
I no longer worry much about calibration issues, but I might try calibrating my laptop with my Eye-one Display 2, just for kicks.
I think I've reached the stage where my aging CRT monior is totally compatible with my Eye-One calibration package.
-
I no longer worry much about calibration issues, but I might try calibrating my laptop with my Eye-one Display 2, just for kicks.
[a href=\"index.php?act=findpost&pid=108790\"][{POST_SNAPBACK}][/a]
I know I do. This of course isn't an ideal display for color critical work but I still find calibrating and profiling such a device has benefits. On the road, when working in Lightroom, I don't have any issues doing work on my raws in such a display, certainly using Quick Develop on lots of images. Since this is metadata editing, I'm never touching the raws. Even if the corrections are off a bit, based on such a display, it gets them in good shape for review and editing down the picks. I can always tweak later on a better display.
-
I know I do. This of course isn't an ideal display for color critical work but I still find calibrating and profiling such a device has benefits. On the road, when working in Lightroom, I don't have any issues doing work on my raws in such a display, certainly using Quick Develop on lots of images. Since this is metadata editing, I'm never touching the raws. Even if the corrections are off a bit, based on such a display, it gets them in good shape for review and editing down the picks. I can always tweak later on a better display.
[a href=\"index.php?act=findpost&pid=108792\"][{POST_SNAPBACK}][/a]
Who was it said that marketing laptop calibration is in large part about managing user expectations ?
Edmund
-
Who was it said that marketing laptop calibration is in large part about managing user expectations ?
Edmund
[a href=\"index.php?act=findpost&pid=108864\"][{POST_SNAPBACK}][/a]
I don't know, who?
-
I don't know, who?
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=108879\")
Maybe I misquoted Tom Lianza ?
"The biggest challenge is the laptop display. These displays are
generally unusable for critical viewing (Mac or PC) and customers insist
on calibrating them. The management of customer expectations in this
area is a full time job. It seems that the displays in this area are
getting worse, not better. "
[a href=\"http://lists.apple.com/archives/colorsync-users/2006/Jun/msg00183.html]http://lists.apple.com/archives/colorsync-...n/msg00183.html[/url]
Edmund
-
Maybe I misquoted Tom Lianza ?
I don't recall that specifically but Tom is one of the top guys in his field.
I don't expect calibrating such displays to make them into Sony Artisans but I find it helps (well it doesn't hurt). Same with projectors. Depending on how old they are and the quality, I'm often amazed at the before and after previews I see when I use a Beamer before doing a presentation. Certainly worth the 5 minute effort.
-
I don't recall that specifically but Tom is one of the top guys in his field.
I don't expect calibrating such displays to make them into Sony Artisans but I find it helps (well it doesn't hurt). Same with projectors. Depending on how old they are and the quality, I'm often amazed at the before and after previews I see when I use a Beamer before doing a presentation. Certainly worth the 5 minute effort.
[a href=\"index.php?act=findpost&pid=108887\"][{POST_SNAPBACK}][/a]
Never calibrated a beamer yet - must try it. I think there is a real calibration cottage industry growing up in the home theatre area.
Edmund
-
I think there is a real calibration cottage industry growing up in the home theatre area.[a href=\"index.php?act=findpost&pid=108892\"][{POST_SNAPBACK}][/a]
Absolutely, although the operative words might be "pleasant" and "realistic" rather than "accurate"
-
How is it these types of "done to death" topics still attract so many views? I mean we're at 1878. What's going on here? The long troubleshooting session with the MacBook Pro titled "OS X color management" is up to 3592 last I checked. Nothing much got accomplished.
Who reads these posts? Marketers? Consumers? Or is it just the "made you look" compulsion in play.
Heck, I'm still here. I guess that's what it is.
-
I mean we're at 1878. What's going on here?
[a href=\"index.php?act=findpost&pid=108928\"][{POST_SNAPBACK}][/a]
Maybe the numbers are not 'accuate' <g>
-
In Texas that's pronounced "aker-uht".
-
I subscribe to the Critical Mass Theory - once a thread reaches it, it becomes a Forum Black Hole, and anyone reading the title basically touches the event horizon and gets sucked in, usually with unwilling yet unmitigated curiosity ...
-
We read these threads because these are basic issues that all users, from the prosumer level on up, are still fuzzy about: the main precepts of color management and digital capture.
Unfortunately what we get when we read, beyond the childish remarks, are industry "pros" (those with the time, inclination, and I assume training to write books and start websites) who can't agree on basic definitions of what they're talking about- we're not talking about PS "different ways of doing different things", we are talking about what should be basic information that everyone should agree upon by now. You would think.
So, like the rest, I'll read, glean what I can, check that against other information, against my own tests and common sense, and make an evaluation. As to whether the earth is round or flat.
-
As to whether the earth is round or flat.
[a href=\"index.php?act=findpost&pid=109242\"][{POST_SNAPBACK}][/a]
I say it's a hecatohedron.