Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: JessicaLuchesi on January 13, 2010, 06:42:13 am
-
Hey everyone,
I have a pretty simple question here, I am reading several comments that the Spyder 3 package is not calibrated from factory ( which sounds so stupid, I would doubt they would actually do this ), and also, fails to properly calibrate the Apple LED displays (iMac / MacBook Pro).
So, you start to second guess the color accuracy of your monitor for even working with it. So, here goes the 1 Million Dollar Question:
Would you think swapping from the Spyder 3 Pro to the i1Display 2 is something you would do? You consider the Spyder 3 to be so unreliable, the profiles it generates are just unusable? Or is it just with the Apple displays?
Since I don't have any other system to compare here, I really have no way of knowing how accurate my own Spyder is.
Will be waiting for replies, and thank you so much for any help, tips, experience and suggestions.
Jessica
-
Do you have a Spyder3 Pro or Spyder 3 Elite? And how is your gray balance?
-
Do you have a Spyder3 Pro or Spyder 3 Elite? And how is your gray balance?
Pro. How do I measure the gray balance using the Spyder?
-
Would you think swapping from the Spyder 3 Pro to the i1Display 2 is something you would do?
First of all, read about the accuracy of low-cost colorimeters: http://lists.apple.com/archives/colorsync-...v/msg00175.html (http://lists.apple.com/archives/colorsync-users/2009/Nov/msg00175.html)
I have some colorimeters (and spectrometers). But only NEC branded X-Rite i1Dsiplay 2 (optimized by NEC for NEC monitors) is perfect for calibrating NEC SpectraView monitors. Other monitors are calibrated with spectrometers, because my "generic" Spyder3 and i1Display 2 are inaccurate, especially Spyder3.
-
First of all, read about the accuracy of low-cost colorimeters: http://lists.apple.com/archives/colorsync-...v/msg00175.html (http://lists.apple.com/archives/colorsync-users/2009/Nov/msg00175.html)
I have some colorimeters (and spectrometers). But only NEC branded X-Rite i1Dsiplay 2 (optimized by NEC for NEC monitors) is perfect for calibrating NEC SpectraView monitors. Other monitors are calibrated with spectrometers, because my "generic" Spyder3 and i1Display 2 are inaccurate, especially Spyder3.
Thank you, I'll be reading in a while. Actually, I did some testings on my Color Checker Passport, and have some questions to post in a while ( will create a new thread for that, want to do some proper lighting and further testing before I do so ).
-
From what I take, the basic differences between Spyder 3 Pro and Elite, is within the software package, not really regarding the sensor itself. And, the software can be upgraded ( or I can buy a third party solution that works with the Spyder hardware ).
So, how could I evaluate if I have a trustworthy sensor in my hands?
-
Hi,
The problem with colorimeters is that they don't measure spectrum but just three colors. So they measure Red, Green and Blue. Problem is that they don't know which what Red, Green and Blue they measure. With cathodic ray tubes it was easy, because they used to have well defined "phosphors" emitting well know wavelengths. With LCDs the situation is a bit different. They have an electroluminescent background light which has some spectral characteristics and than a set of filters (for red, green and blue pixels) that also have spectral properties. A spectrometer measures a spectrum so it can take all those variations into account.
A colorimeter can be matched to a certain screen by choosing suitable filters and having a correct calibration matrix for the screen in question. You may check this article:
http://www.lumita.com/site_media/work/whit...xrite-wp-3a.pdf (http://www.lumita.com/site_media/work/whitepapers/files/xrite-wp-3a.pdf)
The article may be a bit old. Modern display technology may have changed the playing field a bit.
Best regards
Erik
Hey everyone,
I have a pretty simple question here, I am reading several comments that the Spyder 3 package is not calibrated from factory ( which sounds so stupid, I would doubt they would actually do this ), and also, fails to properly calibrate the Apple LED displays (iMac / MacBook Pro).
So, you start to second guess the color accuracy of your monitor for even working with it. So, here goes the 1 Million Dollar Question:
Would you think swapping from the Spyder 3 Pro to the i1Display 2 is something you would do? You consider the Spyder 3 to be so unreliable, the profiles it generates are just unusable? Or is it just with the Apple displays?
Since I don't have any other system to compare here, I really have no way of knowing how accurate my own Spyder is.
Will be waiting for replies, and thank you so much for any help, tips, experience and suggestions.
Jessica
-
From what I take, the basic differences between Spyder 3 Pro and Elite
Datacolor had a manufacturing problem witha batch of S3Pros that can lead to visual discoloration in the graybalance that anyone with a good eye can see. If you have one of these problematic Pros they'll replace them without question. All of the Elite devices have been excellent. Between my own colorimeters and my clients, I get my hands on a lot of these. The EyeOneDisplay has a long history of quality control issues - I'd stay away from it. For demanding environments I think it's impossible to beat Color Eyes Display Pro with a Spyder 3 Elite. FWIW, I wrote a brief article a year ago talking about some of the differences between colorimeters and spectrophotometers for display calibration: http://www.on-sight.com/2008/12/08/updated...ation-packages/ (http://www.on-sight.com/2008/12/08/updated-review-of-colorimeters-and-display-calibration-packages/)
-
This debate between the Spyder 3 and the i1 is getting a bit old. Everyone seems to have an opinion and states it as a truth. I'm coming to the conclusion both are perfectly good devices.
I'd really like to hear a response from the experts here on the OP's direct question... "So, how could I evaluate if I have a trustworthy sensor in my hands?"
I was recently in the same situation of not trusting my device and could not get an answer to this question either. I'm guessing that in an isolated environment with no other results to compare to there is no way to come to a conclusion. I use the Color Eyes Display Pro software and erratic validation results were my only clue.
BTW I got the i1 based on the reports of Spyders not being mfr checked... FWIW
-
Datacolor had a manufacturing problem witha batch of S3Pros that can lead to visual discoloration in the graybalance that anyone with a good eye can see. If you have one of these problematic Pros they'll replace them without question. All of the Elite devices have been excellent. Between my own colorimeters and my clients, I get my hands on a lot of these. The EyeOneDisplay has a long history of quality control issues - I'd stay away from it. For demanding environments I think it's impossible to beat Color Eyes Display Pro with a Spyder 3 Elite. FWIW, I wrote a brief article a year ago talking about some of the differences between colorimeters and spectrophotometers for display calibration: http://www.on-sight.com/2008/12/08/updated...ation-packages/ (http://www.on-sight.com/2008/12/08/updated-review-of-colorimeters-and-display-calibration-packages/)
Scott,
I am new to Color Management, and don't know what you meant by Gray Balance. But, I have a fairly good eye, and a Color Checker Passport, which I may use to compare the patches with the desired RGB counterpart of the 24-patch on screen ( using the data supplied on the Real World Color Management - Second Edition book - Page 250 ), you think it's a silly endeavor or worth a shot?
I just installed Color Eyes Display Pro, and yes, there is a HUGE difference in the results. The "glare" effect of my screen is gone, which was pretty annoying, I always felt the display was too bright. If the devices on the Spyder3 package are the same, as Color Vision claims, maybe that's the route for me, at least, before I can afford a good spectrophotometer (i1Pro I guess).
So, any tips on how someone new to Color Management can check their device for some precision?
Jessica
-
Btw, I have the "validation results" from Color Eyes Display Pro. Could that be helpful to spot problems?
ColorEyes Display Validation Log
2010-01-14 11:47:02
Display: Color LCD 1
System Profile: ColorEyesDemo
Calibrated By: ColorEyes Display
Calibration Date: 110-00-14 11:46:01
Validation Results:
# C M Y K Target L a b Measured X Y Z L a b DeltaE
1 0 0 0 0 100.000 0.000 0.000 164.48 173.03 189.13 100.000 0.017 -0.242 0.26
2 0 0 0 0 94.910 0.000 0.000 143.07 150.84 164.56 94.813 -0.330 -0.111 0.50
3 0 0 0 0 89.822 0.000 0.000 124.33 131.07 142.62 89.742 -0.290 0.050 0.44
4 0 0 0 0 84.745 0.000 0.000 107.64 113.43 123.42 84.769 -0.225 0.054 0.34
5 0 0 0 0 79.660 0.000 0.000 91.48 96.45 105.18 79.467 -0.285 -0.070 0.45
6 0 0 0 0 74.577 0.000 0.000 77.99 82.13 89.01 74.485 -0.105 0.250 0.30
7 0 0 0 0 67.155 0.000 0.000 59.84 62.95 68.74 66.808 0.033 -0.131 0.31
8 0 0 0 0 58.956 0.000 0.000 43.92 46.29 50.20 58.742 -0.178 0.178 0.37
9 0 0 0 0 50.393 0.000 0.000 30.84 32.46 35.42 50.403 -0.040 -0.079 0.10
10 0 0 0 0 37.975 0.000 0.000 16.47 17.39 18.82 37.933 -0.283 0.195 0.46
11 0 0 0 0 25.684 0.000 0.000 7.46 7.87 8.60 25.400 -0.154 -0.090 0.32
12 0 0 0 0 17.782 0.000 0.000 4.01 4.23 4.66 17.677 -0.147 -0.217 0.32
13 0 0 0 0 56.206 82.374 80.182 75.84 39.97 2.82 56.321 82.101 80.676 0.27
14 0 0 0 0 86.791 -78.936 91.187 58.56 119.43 18.24 86.715 -79.061 91.237 0.06
15 0 0 0 0 30.162 60.827 -94.913 29.80 13.02 167.54 30.254 60.802 -94.816 0.08
Maximum DeltaE: 0.50
Average DeltaE: 0.31
-
So, any tips on how someone new to Color Management can check their device for some precision?
It is hard to do without a reference sensor. Even for professionals.
-
It is hard to do without a reference sensor. Even for professionals.
All I know, is this MacBook Pro screen is kicking my a$$, LOL
Opening the reference 24-patch ProPhoto file ( http://www.colorremedies.com/realworldcolor/downloads.html (http://www.colorremedies.com/realworldcolor/downloads.html) ) and comparing to the X-Rite color checker passport 24-patch target, and looking one by one, it seems, on visual inspection, the color are spot on, but, very very "bright". I mean, considering the target is validated and accurate ( which X-Rite certifies for two years ), what I felt is, there was no visual shift in the colors, but, they were washed off on the screen, compared to the target. Even configured the Color Eyes Display Pro to a White Luminance Target of 110cd/m2, and it still looks washed off ( not sure if the software could come down to 110cd/m2 ).
Being on a budget sucks, otherwise I could just order an Eizo Monitor and i1Pro and be over with this :|
-
Hi!
Did you measure your brightness? Can you reduce brightness in your display settings? Ideally, brightness should be a like a white sheet of paper in your reference illumination. So if your screen is brighter than a sheet of paper in the illumination you use to evaluate your prints it would be too bright.
Best regards
Erik
All I know, is this MacBook Pro screen is kicking my a$$, LOL
Opening the reference 24-patch ProPhoto file ( http://www.colorremedies.com/realworldcolor/downloads.html (http://www.colorremedies.com/realworldcolor/downloads.html) ) and comparing to the X-Rite color checker passport 24-patch target, and looking one by one, it seems, on visual inspection, the color are spot on, but, very very "bright". I mean, considering the target is validated and accurate ( which X-Rite certifies for two years ), what I felt is, there was no visual shift in the colors, but, they were washed off on the screen, compared to the target. Even configured the Color Eyes Display Pro to a White Luminance Target of 110cd/m2, and it still looks washed off ( not sure if the software could come down to 110cd/m2 ).
Being on a budget sucks, otherwise I could just order an Eizo Monitor and i1Pro and be over with this :|
-
Btw, I have the "validation results" from Color Eyes Display Pro. Could that be helpful to spot problems?
Maybe, but probably not since if the fault or inaccuracy is in the instrument, how can it report this inaccuracy when you use the same device as a double check? Its like having a 12” ruler that’s off by ⅓ of an inch. You can measure the same thing a dozen times and its still ⅓ inch off and worse, not aware of this inaccuracy. That and the fact that the colors used for such reports are not necessarily the most demanding colors to hit in color space (the manufacturer gets to pick what colors give us the report).
-
...Its like having a 12” ruler that’s off by ⅓ of an inch. You can measure the same thing a dozen times and its still ⅓ inch off and worse, not aware of this inaccuracy....
Yeah, but it can demonstrate consistency. I used to have a Spyder3 Pro puck and it could validate a ColorEyes created profile 3 times with 3 widely and wildly varying results. I replaced it with a Spyder3 Express puck and now 3 consecutive validations will give consistent results. I can't attest to accuracy, but at least it shows it's consistent.
-
Yeah, but it can demonstrate consistency.
Yes, and trending is very useful. That’s what makes Maxwell so cool (http://www2.chromix.com/maxwell/index3.cxsa). As long as its understood there can be consistent incorrect measurements.
-
Yeah, but it can demonstrate consistency. I used to have a Spyder3 Pro puck and it could validate a ColorEyes created profile 3 times with 3 widely and wildly varying results. I replaced it with a Spyder3 Express puck and now 3 consecutive validations will give consistent results. I can't attest to accuracy, but at least it shows it's consistent.
Good idea. Am gonna give it a try, if I see much deviation, well, I'm adding an i1 Display 2 to my shopping cart. And then save to an i1Pro ( and maybe an Eizo monitor if this is a pretty good year ).
-
Quick note, opening the ProPhoto and LAB versions of the 24-patch target from the book's support site had some minute differences when comparing my own screen to the physical X-rite target. The LAB file was almost spot on. The only issue still to be looked upon, is the brightness, not really color matching. The bottom gray patches still seem to be really bright, compared to the card. It's also noticeable on the color patches, but it's impossible to ignore on the gray ones.
Anyone has any idea how to solve this? Maybe changing the target white or black luminance while calibrating?
The ColorEyes Display Pro suggests 120cd/m2, but I'm setting to 100cd/m2 and it still seems too bright. Is there some sort of "standard" value to be used?
-
Hi!
Did you measure your brightness? Can you reduce brightness in your display settings? Ideally, brightness should be a like a white sheet of paper in your reference illumination. So if your screen is brighter than a sheet of paper in the illumination you use to evaluate your prints it would be too bright.
Best regards
Erik
Hey Erik,
No, I didn't. How can I measure it using either the Spyder3 Pro software, or the ColorEyes package? I did however set, and try to vary, the target luminance while calibrating. I can manually set it from the system preferences, yes. I tried dimming, but I didn't feel I was getting any closer to solving the problem. I think, since the ColorEyes software adjusts everything on the LUT ( from what I inferred ), I should try to set it there. Not sure if I am correct. Maybe I am indeed running after my own tail, and getting just slightly closer, but without any definite solutions
Thank you for the tip on the printing evaluation, but I really don't print. My files go directly to editors, and even they don't have perfect proofing of the prints. They rely on experience to sort out problems. I think my best chance, is to maybe match a patch image to the GretagMacbeth 24-patch target, which is a known, tested reference. Still, I guess it's only as good and accurate as my eyes. And even then, not really an official image, but one provided as support material to the Real World Color Management book.
-
Hi,
Yes, the Color should adjust the LUT so that you get the right brightness. I'm using a Color Munki, and with that tool there is a step in the calibration process where you can adjust brightness.
As a sanity check you may check out some web page for visual profiling of screens like this one:
http://www.normankoren.com/makingfineprint...html#gammachart (http://www.normankoren.com/makingfineprints1A.html#gammachart)
On thing you could try is to take a some standard image (like the image on Andrew Rodney's web page: http://www.digitaldog.net/files/Printer_Test_file.jpg.zip (http://www.digitaldog.net/files/Printer_Test_file.jpg.zip) ), process to your taste and post it
on this forum. This should indicate if you have a serious issue.
Regarding "gamma" I'd use 2.2 regardless of you having a Mac, another alternative is to use "Native gamma".
Sorry I cannot help!
Best regards
Erik
Hey Erik,
No, I didn't. How can I measure it using either the Spyder3 Pro software, or the ColorEyes package? I did however set, and try to vary, the target luminance while calibrating. I can manually set it from the system preferences, yes. I tried dimming, but I didn't feel I was getting any closer to solving the problem. I think, since the ColorEyes software adjusts everything on the LUT ( from what I inferred ), I should try to set it there. Not sure if I am correct. Maybe I am indeed running after my own tail, and getting just slightly closer, but without any definite solutions
Thank you for the tip on the printing evaluation, but I really don't print. My files go directly to editors, and even they don't have perfect proofing of the prints. They rely on experience to sort out problems. I think my best chance, is to maybe match a patch image to the GretagMacbeth 24-patch target, which is a known, tested reference. Still, I guess it's only as good and accurate as my eyes. And even then, not really an official image, but one provided as support material to the Real World Color Management book.
-
Hi,
One thing that you should keep in mind that the black box on the Macbeth card is not black, it's more like dark grey. You could check out this page for black levels:
http://www.normankoren.com/makingfineprint...html#gammachart (http://www.normankoren.com/makingfineprints1A.html#gammachart)
I don't know about black luminance, because I don't have it my tool set. Regarding white luminance it's really depends on your viewing environment. The 100 cd/m2 seems like a reasonable value to me.
Best regards
Erik
Quick note, opening the ProPhoto and LAB versions of the 24-patch target from the book's support site had some minute differences when comparing my own screen to the physical X-rite target. The LAB file was almost spot on. The only issue still to be looked upon, is the brightness, not really color matching. The bottom gray patches still seem to be really bright, compared to the card. It's also noticeable on the color patches, but it's impossible to ignore on the gray ones.
Anyone has any idea how to solve this? Maybe changing the target white or black luminance while calibrating?
The ColorEyes Display Pro suggests 120cd/m2, but I'm setting to 100cd/m2 and it still seems too bright. Is there some sort of "standard" value to be used?
-
A while back I created a color target for assessing display calibration accuracy especially for gamma and color matrix formula transform in color managed apps.
These colors were derived from years of noticing their hue/saturation response shift in Photoshop on a Mac system when loading different display profiles that used inaccurate color matrices encoded within their build.
Some of these matrices=(mathematical descriptors of combined display colorant purity's and color temp) are embedded within the default profile built by the system whenever a new display is connected and not profiled. Most displays somewhere in the late '90's had their color responses measured and written into their ROM chip by the factory in standard EDID format. Often they either were inaccurate or were written in such a way that induced the OS to create a corrupt default profile causing hue/saturation shifts in Photoshop.
Mac's eyeball calibrator in OS X automatically derives these EDID numbers whenever you see "Color LCD" or the manufacturer's brand name in a profile the user didn't install.
Anyway, here's what to visually test for in this target, it's in the AdobeRGB space so it needs to be viewed in a color managed app:
The small squares in the center of the bigger squares should not be more noticeable in luminance than the differences seen between the solid black square and the gray square right above it except of course for the Golden Yellow square and blue gradient-(my homage to Maxfield Parrish).
The red in the Xrite color chart should not look orange or magenta. The baby on the far right should have a more noticeable amount of yellow in its peachy pink tan complexion compared to the pink baby second from the left. And none of the colors in the entire chart should be glowing (blooming) or appear over saturated.
Hope it helps, cheers!
Tim Lookingbill
[attachment=19478:ColorSqu...inTarget.jpg]
-
A while back I created a color target for assessing display calibration accuracy especially for gamma and color matrix formula transform in color managed apps.
These colors were derived from years of noticing their hue/saturation response shift in Photoshop on a Mac system when loading different display profiles that used inaccurate color matrices encoded within their build.
Some of these matrices=(mathematical descriptors of combined display colorant purity's and color temp) are embedded within the default profile built by the system whenever a new display is connected and not profiled. Most displays somewhere in the late '90's had their color responses measured and written into their ROM chip by the factory in standard EDID format. Often they either were inaccurate or were written in such a way that induced the OS to create a corrupt default profile causing hue/saturation shifts in Photoshop.
Mac's eyeball calibrator in OS X automatically derives these EDID numbers whenever you see "Color LCD" or the manufacturer's brand name in a profile the user didn't install.
Anyway, here's what to visually test for in this target, it's in the AdobeRGB space so it needs to be viewed in a color managed app:
The small squares in the center of the bigger squares should not be more noticeable in luminance than the differences seen between the solid black square and the gray square right above it except of course for the Golden Yellow square and blue gradient-(my homage to Maxfield Parrish).
The red in the Xrite color chart should not look orange or magenta. The baby on the far right should have a more noticeable amount of yellow in its peachy pink tan complexion compared to the pink baby second from the left. And none of the colors in the entire chart should be glowing (blooming) or appear over saturated.
Hope it helps, cheers!
Tim Lookingbill
[attachment=19478:ColorSqu...inTarget.jpg]
Thanks Tim,
I also did something similar myself. Since I do own a ColorChecker Passport, I went the dumb way. If the purpose of that target and software is to generate a proper camera profile, then, by properly lighting it and metering it, that image, once white balanced and profiled, should give an image of the target comparable to the target itself.
The several online target-images available looked "washed out" for me, and I was thinking either my screen or my Spyder had a problem. But both your photo and my own image of the target ( with the DNG profile created by the CC Passport plugin ) were spot on. My own image a bit more than yours, but both were more accurate to the actual physical target ( placed side by side with the screen ) than the downloaded patches. But even the downloaded patches didn't have color changes, as I could see the same colors, without even slight variations, it just looked "washed out".
I also did the test suggested on "Real World Color Management", page 220 ( Chapter 9 - Figure 9-1 ), which is their Black Point Check. They state an excellent calibration system should see a difference between level 0 and 1, and a typical system, around 5 or 7. I noticed a change right on the first level, without color casting.
I guess that leaves me with some peace of mind. Yes, it's a notebook screen, it will never be perfect, but at least my Spyder doesn't seem to be giving me false colors, or making images so bright, I could get an exposure wrong, which were my biggest fears. I still want to go towards a i1Pro and good external monitor when my budget allows. But I don't think I need to replace my colorimeter right now. Maybe I got lucky and got one of the "good" Spyder3 units.
Anyway, it was also a couple of days of testing every possible white luminance, black luminance and recalibrating and comparing the CC Passport to images of the target, trying to find significant shifts, aberrations, and in the end all the hard work turned out fine
Thanks everyone, I really mean it, all the help in this forum was really important, even to wonder what else I could check and do and test and see how to improve.
And here's what I did to manage a decent calibration on my MacBook Pro 13" screen.
Software: ColorEyes Display Pro
Hardware: Spyder3
Monitor Settings: Apple Cinema Display,iMac, Laptop
Profile Settings: ICC v4
White Point Target: D65, 110cd/m2
Gamma: 2.2
Black Point Target: 0.35cd/m2
I guess it helps. By the way, the native white luminance of my display is around 195cd/m2. And I did add a few custom gray balance points ( 138, 114, 86, 74, 54, 36 ), because I could see some banding forming on the scale.
-
Put aside the unit to unit inconsistency of spyder 3 pro, the PRO software is simply not good enough because of the manual setting of the target, calibration. ELITE softWare is a must for user who expect good softproofing result. The boiled down most important brench mark of one's system in respect of color management.
In short if you are spared of the color shift issue of you spyder3 hardware, please upgrade to elite. The upgrade price is just the price difference between Pro and Elite , so you won't regret not to buy Elite in the place.
-
Oops
I mean ....because of the manual setting of the target is missing...
-
Oops
I mean ....because of the manual setting of the target is missing...
Yes, but I solved that by using the ColorEyes Display Pro software, instead of the Spyder Pro software. Worked like a charm
-
Yes, but I solved that by using the ColorEyes Display Pro software, instead of the Spyder Pro software. Worked like a charm
Glad to hear it. CEDP has been a strong solution for demanding professionals for quite some time now. From my observation, people often start with a Huey or EyeOneDisplay but eventually make their way up to CEDP. Their iterative calibration process and white point tuning features are key.