Ad
Ad
Ad
Pages: « 1 2 3 [4]   Bottom of Page
Print
Author Topic: Dual Eizo Calibration: one pinkish, one neutral  (Read 12698 times)
Czornyj
Sr. Member
****
Offline Offline

Posts: 1371



WWW
« Reply #60 on: June 11, 2010, 06:20:01 AM »
ReplyReply

Quote from: shayaweiss
John,

These are indeed very useful ratings. My photographic supplier advertised them as suitable for photographic work, but now that I had a closer look at Eizo's Website, I saw that they do also state that it is suitable for that purpose, but only in a "prosumer" or amateur setting, though I cannot understand what they should do with it.

Unfortunately, Eizo does have a distributor here who sells their screens double the price (yes!) from Europe, even for models that have already been replaced, like the CD222W. That's why, when I paid for the Flexscan s2233w, I paid like you would pay for the ColorEdge CD222w. I thought I'll try to safe some money. It was a big mistake.

Take a look at NEC displays - the new PA241W is one of the finest displays I've ever saw, I dare to say it's at least as good as Eizo CG243W (or maybe even better). It's perfectly uniform and linear, and very flexible.
Logged

shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #61 on: June 11, 2010, 06:30:32 AM »
ReplyReply

Quote from: Czornyj
It's not i1d2, it's only my DTP94 (that's even better than i1). NEC US makes his custom matted colorimeters for a good reason. I'm not trying to split hairs, I'm only trying to explain what could be the reason of the problem.

My picture is only an example, and i1pro cannot be considered as a serious source of reference measurement values, but here's the result of a serious scientific test:
http://luminous-landscape.com/forum/index....mp;#entry324899

I found this document very interesting, Czornyj! I looked it up in the original thread on colorsync http://lists.apple.com/archives/colorsync-...v/msg00175.html .
« Last Edit: June 11, 2010, 06:33:23 AM by shayaweiss » Logged
John R Smith
Sr. Member
****
Offline Offline

Posts: 1357


Still crazy, after all these years


« Reply #62 on: June 11, 2010, 08:17:55 AM »
ReplyReply

shaya

Steve has been endlessly patient and has gone beyond the call of duty to help you out with this problem, but personally I think it may have very little to do with the Win 7 setup or the drivers or the graphics card, but probably everything to do with the Eizo panels themselves. You have exposed their weaknesses precisely by setting them up in a dual monitor situation where you can see the inconsistency between samples.

I do know a bit about this, because in a general sense I look after about 35 PCs and ten laptops for our service. In our room there are 6 PCs which have profiled monitors, for which I use a Spyder 2. One of the monitors is an Eizo CG211, which is an absolute peach - with perfectly neutral luminance over the whole panel. The rest are cheap Philips 190s, but they profile just fine - until you try putting two side by side! Then you will see subtle hue shifts which you would never have noticed otherwise. An equally cheap Samsung was a total disaster, with a green cast to the supposedly neutral grays which I could not profile out.

John
Logged

Hasselblad 500 C/M, SWC and CFV-39 DB
and a case full of (very old) lenses and other bits
shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #63 on: June 11, 2010, 09:05:59 AM »
ReplyReply

Quote from: John R Smith
shaya

Steve has been endlessly patient and has gone beyond the call of duty to help you out with this problem, but personally I think it may have very little to do with the Win 7 setup or the drivers or the graphics card, but probably everything to do with the Eizo panels themselves. You have exposed their weaknesses precisely by setting them up in a dual monitor situation where you can see the inconsistency between samples.

I do know a bit about this, because in a general sense I look after about 35 PCs and ten laptops for our service. In our room there are 6 PCs which have profiled monitors, for which I use a Spyder 2. One of the monitors is an Eizo CG211, which is an absolute peach - with perfectly neutral luminance over the whole panel. The rest are cheap Philips 190s, but they profile just fine - until you try putting two side by side! Then you will see subtle hue shifts which you would never have noticed otherwise. An equally cheap Samsung was a total disaster, with a green cast to the supposedly neutral grays which I could not profile out.

John

Steve and John,

Steve, you have indeed been very patient with me and I was so much hoping that we'll work it out. And maybe, as you suggested, we could still try to reinstall my whole system. But as you said, Steve, reality makes that I can't afford to do so.

On one hand, the problems with the panels (that are made by Samsung I heard, right?) also seems to have revealed a weakness in in the i12 and Match3, non? Why should this be due to the monitors? As you said, Steve, in majority of cases this combination might work well, but if there is a problem with the panel, like the magenta cast, than this instrument in combination with that software fails. That's how I understood the situation till now. And maybe somehow there could be some driver or other system related problem that creates or accentuates all this.

The absence of uniformity in brightness and chroma, on the other hand, is certainly not related to any software problem. A calibration with a different software, iColor Display 3, was able make things a little bit better.
« Last Edit: June 11, 2010, 09:11:28 AM by shayaweiss » Logged
Steve Weldon
Sr. Member
****
Offline Offline

Posts: 1445



WWW
« Reply #64 on: June 12, 2010, 12:03:22 PM »
ReplyReply

Quote from: shayaweiss
Steve and John,

Steve, you have indeed been very patient with me and I was so much hoping that we'll work it out. And maybe, as you suggested, we could still try to reinstall my whole system. But as you said, Steve, reality makes that I can't afford to do so.

On one hand, the problems with the panels (that are made by Samsung I heard, right?) also seems to have revealed a weakness in in the i12 and Match3, non? Why should this be due to the monitors? As you said, Steve, in majority of cases this combination might work well, but if there is a problem with the panel, like the magenta cast, than this instrument in combination with that software fails. That's how I understood the situation till now. And maybe somehow there could be some driver or other system related problem that creates or accentuates all this.

The absence of uniformity in brightness and chroma, on the other hand, is certainly not related to any software problem. A calibration with a different software, iColor Display 3, was able make things a little bit better.
Like I said, you probably won't find it worth it.  It depends on if your goal is to learn exactly what the problem is.. or just to get your workstation going.  I don't blame you either way.  Myself, I want to understand what's going on because there will always be a 'next time' and with so much work already invested it seems a shame to short change the learning opportunity.

You also don't need to reinstall your entire system.  You can install a copy of Win7 and the necessary drivers on another hard disk or even another partition while keeping your current system intact.

And no.. I don't think you've revealed a weakness in anything at this point.  The best you can do is stick your finger in the wind and guess.  Everything is a guess until you find the actual issue at hand.  What you find might or might not help others.. or yourself.

If you're now properly profiling using the 1i2 and different software.. then it's not the 1i2 causing YOUR issues.. indeed it might have variance as has been described.. but chances are that variance will hold steady across two monitors on the same system and render equal variances from the norm.. but it's rather obvious it's not the hardware causing your particular issue.. because you're using that hardware with different software to good effect.

It's not necessarily the software either.. it could be any number of system issues and left over whatever's from all you're experiments interfering or bugging that Imatch3 software.. but that doesn't make the software bad.  That makes it an operator induced issue.  It could be the software.. we just don't know that with enough confidence to say it is.

And it's not Win7 because others are using Win7 to profile dual monitors and not getting magenta casts.

Without your third monitor to test you can't say it's that either.  A few days ago you thought all three were bad.. just because one was worse than the other doesn't mean it wouldn't correct equally as well if properly calibrated.

So no.. we don't know.  We can guess.. but from my point of view every piece of software and hardware in question is working fine in other peoples systems without pink casts.  I'm leaning towards something in your system, brought on from frequent and haphazard installs and deinstalls, being the real problem.  Every time you uninstall software.. especially drivers.. you leave behind a lot of things you don't realize.

Now here's the question.. Lets say I'm right.. and left behind in your system is a registry setting, file remnant, whatever.. which is causing the problem.. Currently you've bypassed the issue by using a piece of software that accesses different registers, different registry settings, etc.  Fine.  But what else are these issues interfering with?  Print profiles maybe?  CS5?  Who knows..   And if not anything now, what will they interfere with in the future.. and how much more time will you spend figuring out a way around that problem?

It would take less than an hour to set up Win7 and the proper drivers and software on a spare HDD or different partition (I recommend an entire different hard drive.. or at least 'image' your current system drive before proceeding)..  If it works as it should we learned its not certain things.. and I think that would be enough.  Finding the actual remnant or registry setting and what causes it is of less value because it would take a certain sequence of events to duplicate.. knowing what works in the hardware and software department.. is valuable.

Regardless.. I'm really glad you got it going.
Logged

----------------------------------------------
http://www.BangkokImages.com
Steve Weldon
Sr. Member
****
Offline Offline

Posts: 1445



WWW
« Reply #65 on: June 12, 2010, 12:20:00 PM »
ReplyReply

Separate issue along the same topic:

A few months back I ordered (2) NEC LCD2690uxi2's to be used in a dual monitor config.  I've been using dual monitor configs for a decade and have set up at least a few dozen for clients.  A few things I've learned.

1.  If you want a true visual match, side by side, then get two identical monitors of the exact same vintage.

2.  Make sure they're hooded.  Reflected colors from walls/ceilings/table tops (even a colorful binder on the table) can make you perceive colors differently on one monitor and not another.  My table tops are 18% grey (seriously), my walls a neutral white, and what lighting I used is planned and uniform temps.

3.  Use the exact same cables, make sure they don't run parallel to power cables or any other cables.  Issues from this are very rare but they do happen.

4.  If you get a perfect match say at 5000/2/110.. you might not get a match at 6500/2.2/170.. or any other setting.  Monitors perform differently at different whitepoints and luminance values.

5.  You need to train yourself to 'see' to eliminate casts.  What I mean.. is before looking at the white screens for a visual match make sure you've been viewing a neutral color and you shift your eyes back and forth in the same way every time.  How the eye perceives color is greatly influenced by what they perceived immediately before.


There's more.. but these are the bigger issues I've noticed.

With my (2) NEC LCD2690uxi2's I've had a new issue related to number 1 above.  One monitor arrived 30 days before the other.  By the time the second arrived I had about 500 hours on it.  I could get my 1i2 and SVII (software) to profile them exactly.. and the colors looked matched.. but the white screens did not.  They were rendering a different visual whitepoint.  This is common enough for NEC to mention the issue and recommend a 'fix' in their SVII manual.

I was easily able to perform the fix for all color profile types I use, except for the sRGB Emulation.  Why?  Because the only variable NEC allows you to adjust on the sRGB emulation is the lum.. You also can't build (or I couldn't figure out how to build) a sRGB emulation mode that limits the gamut to sRGB like the original sRGB Emulation modes do..

This was driving me nuts.. being an anal type the different visual white point really bothered me.  I'm running the sRGB Emulation profiles at 175 vs. the original unlimited.  And I couldn't fix it.  So.. I marked a day on my calender where I'd be available to make the call to NEC for some help.  Being in Bangkok the time/day differences require some planning.

Between the time I was having the problem, and he time I planned (4 days) my second monitor started matching the first more and more without any change in calibration.  At first I thought my eyes were playing tricks on me as the change was extremely gradual.  Now.. as the monitor has barely 100 hours on it the change is evident and the monitors almost perfectly match.  At this rate in another few days they will match perfectly.

This tells me these monitors 'break in' to an extent.. or more likely the supporting electronics and CCFL's do.. and that number 1 above is more critical than I originally thought.

With all that said.. I'm very happy with my new monitors.
Logged

----------------------------------------------
http://www.BangkokImages.com
John R Smith
Sr. Member
****
Offline Offline

Posts: 1357


Still crazy, after all these years


« Reply #66 on: June 12, 2010, 01:09:24 PM »
ReplyReply

Steve

Apparently Eizo monitors are also noted for having a "break-in" period. I noticed this with our CG211, when I first profiled it, it was spot-on for neutral grays but after a few weeks it had drifted.

Perhaps Shaya might try running a utility like CClean to tidy up the Win 7 Registry?

John
Logged

Hasselblad 500 C/M, SWC and CFV-39 DB
and a case full of (very old) lenses and other bits
Steve Weldon
Sr. Member
****
Offline Offline

Posts: 1445



WWW
« Reply #67 on: June 12, 2010, 01:32:04 PM »
ReplyReply

Quote from: John R Smith
Steve

Apparently Eizo monitors are also noted for having a "break-in" period. I noticed this with our CG211, when I first profiled it, it was spot-on for neutral grays but after a few weeks it had drifted.

Perhaps Shaya might try running a utility like CClean to tidy up the Win 7 Registry?

John
John -

I would guess all monitors/electronics change values slightly as they 'break-in' or age.. and short of special self calibrating circuitry they always will.  This is most likely the main reason most companies recommend a fresh calibration every 2-4 weeks.  My surprise comes from how quickly in their lifespan these particular monitors demonstrated a difference.  It also makes me worry about getting a replacement monitor under warranty repair if the need should ever arise.  I'd also guess that once I get 1000+ hours on these they'll perfectly match and their aging/degradation will stay linear in nature.

Running dual monitors does present some challenges..

Registry cleaners.  Most only remove orphans and really obvious conflicts.  Orphans don't affect the system other than occupying a bit of memory.. and I doubt this issue would show up as an obvious conflict.. but I'm often wrong about such things.. so if he already has a registry cleaner he trusts it might outweigh the inherent risks registry cleaners present.. not sure.

I've had clients tell me of such issues before and once on site once found pink blinds reflecting on one of the two monitors.  Because of the angle of tilt (they should be slightly tilted in vs. side by side) this affected one but not both.  Another time it was a designer lamp shade.. without being on site with the OP's system you can only guess at so much.. but anything that would normally affect your captures in the field (color casts from walls, green lawns, ceilings, and so forth) have the potential to affect your viewing monitors..

When you try to help someone on the internet with such a complex issue.. you never really expect to be right.. but you hope it motivates their thought in a direction that helps..
Logged

----------------------------------------------
http://www.BangkokImages.com
shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #68 on: June 12, 2010, 03:25:19 PM »
ReplyReply

Steve and John,

I regularly run CCleaner and also something called PowerSuite which was recommended to me by Windows support. I try to avoid installing new things. Only now I tried iColor and BasICColor.

I want to learn and understand, that's why I started this thread here (and not somewhere else...) Now I'll have to find the time to install windows on a different hard drive with no extras, boot from there and see what happens, correct? Maybe I could just run win7 in save-mode?

I tried again to run the monitors connected to a different workstation, and their behavior was quite the same on that different computer, where I did not install all this software, except Match3 to make i12 run.

All monitors here are properly hooded. The darkroom has 18% gray walls and furniture including the desktop, glare free JUST ambient light (about 8 lux) with outside light completely blocked out. Because of these issues I have, I frequently checked all ambient light conditions lately.


Logged
shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #69 on: June 20, 2010, 08:27:41 PM »
ReplyReply

Hi,

A week later, having tried and retried everything that was mentioned here, still, they are far apart! Right now, I have one pinkish and the other one greenish.

I also tried Argyll with DTP94b, the two Eizos don't come closer! Argyll and the closest common denominator between the two screens as color temperature, 5800K, helped a lot with uniformity, but...

What did I not try? What bothers me is that I have ABSOLUTELY NO CLUE of what's going on. I would like to understand...

Even the Monitors are not the same anymore, as one was already changed by the vendor. I'll give back a second one in the coming days.

Anyone?
« Last Edit: June 20, 2010, 08:37:16 PM by shayaweiss » Logged
shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #70 on: June 20, 2010, 09:33:32 PM »
ReplyReply

WOW!

I have no Idea why but... YES... They are as close as it can get. Only when I spread one white documet over both, can I sea one pinkish and one greenish tint, but that's normal I think, as the application can only pick up one profile at a time.


I recalibrated and reprofiled, again from scratch, deleting, reseting and uninstalling everey old profile and driver. Than I used i1d2 with Quato's iColor Display 3. And that's final.

The only reason I can figure out, why this works, is that Quato's software has these corrections for wide gamut displays, maybe something close to what the filters do in the custom mated i1d2 by NEC. I did already get the best results with this combination before, but now, it's ALMOST PERFECT. I have still no clue what I did different this time!? I'll think about it, and eventually let you know.
« Last Edit: June 20, 2010, 09:37:33 PM by shayaweiss » Logged
Steve Weldon
Sr. Member
****
Offline Offline

Posts: 1445



WWW
« Reply #71 on: June 20, 2010, 11:42:21 PM »
ReplyReply

Quote from: shayaweiss
WOW!

I have no Idea why but... YES... They are as close as it can get. Only when I spread one white documet over both, can I sea one pinkish and one greenish tint, but that's normal I think, as the application can only pick up one profile at a time.


I recalibrated and reprofiled, again from scratch, deleting, reseting and uninstalling everey old profile and driver. Than I used i1d2 with Quato's iColor Display 3. And that's final.

The only reason I can figure out, why this works, is that Quato's software has these corrections for wide gamut displays, maybe something close to what the filters do in the custom mated i1d2 by NEC. I did already get the best results with this combination before, but now, it's ALMOST PERFECT. I have still no clue what I did different this time!? I'll think about it, and eventually let you know.
Hi Shayaweiss.  I'm glad you found an acceptable match.  Does this mean you won't be getting the NEC's?

If the colors are correct except for the white screen across both monitors then you are as close as you can get without hand adjusting the whitepoint for visual discrepancies.  I think I explained how my new NEC monitors exhibited the same issue with a white screen as they were separated by 500+ hours of run time.. but now that the newest one has hours on it, it's now matching fine.  The age of the monitors indeed is a factor.

In the NEC Spectraview II manual they explain how to adjust for this visual mismatch.  You might find it useful to download the manual and see if you can transfer those procedures over to the software you're using.. I'm not sure if your software supports it.. but give it a read and maybe it will.

I think you're getting there for several reasons.  1.  The video card you're now using supports two LUT's.. and you got rid of that HDMI to DVI adapter..  Two big things.   Deleting all unused drivers and registry entries is also helpful.

I find the tints on the white screen annoying at  best.. which is why I go to great care to buy matched monitors.. but if your colors and grays appear correct then you probably have a useful match.
Logged

----------------------------------------------
http://www.BangkokImages.com
shayaweiss
Jr. Member
**
Offline Offline

Posts: 87


« Reply #72 on: June 21, 2010, 04:46:48 AM »
ReplyReply

I offer following CONCLUSION

[!--quoteo(post=0:date=:name=http://hoech.net/dispcalGUI/)--][div class=\'quotetop\']QUOTE (http://hoech.net/dispcalGUI/)[div class=\'quotemain\'][!--quotec--]A note about colorimeters, wide-gamut displays and dispcalGUI

Colorimeters need a correction matrix in hardware or software to obtain correct measurements from wide-gamut displays. The latter is currently not supported by dispcalGUI, so if you own a wide-gamut display and colorimeter which has not been specifically tuned for this display (i.e. contains a correction matrix in hardware), you will need a spectrometer to accurately measure such a screen.[/quote]

IMHO, that confirms, what Czornyj wrote and my findings, already mentioned earlier. There is one decisive factor. I'll paraphrase the above quoted Florian Hoech: When using a colorimeter with a wide gamut monitor, the device must have a correction matrix applied in hardware, as for the NEC custom mated i1d2 sensor, or applied through the software, like Quato's iColor Display 3. I don't now about the EIZO setup, but it guess ColorNavigator must apply a correction matrix, doesn't it?

This finally explains, why my wide gamut monitors

 - do not work with a regular colorimeter, like the DTP94 or the i1 Display 2 (you might not notice this, as long as you have only one monitor, and the tint is not too strong, as the sensors seems to be consistently off between every characterization and profiling) combined with a software not applying a corrective matrix, like x-rites eye-one Match3 (don't tell me it works, just read the uncountable lamenting threads out there), BasICColor Display 4, or even SpectraView II and Argyll.

 - but they do work with a software matrix filter like Quato's Display 3. (I don't know of any other. Do you?)

 - I could not verify that it should be fine with the matrix filter mated hardware, like the NEC sensor (We are trying to buy one. Any cheap suggestions with overseas delivery?),
 
 - or, as stated above, I could use a spectrometer, as a workaround, having all the disadvantages of the spectrometer measuring a display.


Further questions:

1 ) Why are they consistently off by the same different degree proper to each monitor?
2 ) Why is it that ( and I read this also in a number of desperate user comments ) after characterization, but before profiling the monitor always seems to look a lot better?

Anyone?

P.S.: Evidently, I agree with you, Steve, that changing the graphic card, the cables and connections, as well as cleaning up the drivers etc. is a prerequisite through which you have guided me with a lot of patience. And yes, I will get the new PA241 hopefully soon, because things are now maybe perfect in the world of Flexscan, but still I have not a good enough uniformity in brightness and chroma. It disturbs me and my colleagues here too much.
« Last Edit: June 21, 2010, 06:51:17 AM by shayaweiss » Logged
WombatHorror
Sr. Member
****
Offline Offline

Posts: 299


« Reply #73 on: June 27, 2010, 01:38:28 AM »
ReplyReply

Quote from: Steve Weldon
2.  If you don't have internal LUT's the next step is to call Nvidia and confirm if your card had dual LUT's or not.  I'm almost positive it does not.. but never take anyones word for these things.. and be careful what technician you listen to at Nvidia.. if they go "LUT what?" or hesitate in anyway.. or doesn't know immediately what you're talking about.. then ask to talk to a senior/advanced tech.. because all that guy will do is read the same manual you've already read.. and probably come up with the wrong answer.

Don't most nvidia cards these days have dual LUTs?

All of my recent ones have had that.
Logged
WombatHorror
Sr. Member
****
Offline Offline

Posts: 299


« Reply #74 on: June 27, 2010, 01:51:56 AM »
ReplyReply

Quote from: Steve Weldon
This is great news.. but I must admit I'm more intrigued than ever.  Despite Czornyj's explanation of using a colorimeter, specifically 1i2 and Imatch3, this combination is not only working correctly on tons of wide-gamut monitors.. but are in fact recommended and listed as compatible with wide-gamut monitors.  Yes, there might be some minor differences at the fringes.. but not enough to using anything else if it costs you even a single dollar.

And in fact.. I'm using the 1i2 puck with my wide-gamut (NEC LCD2690uxi2's) monitors with excellent results.

I don't think an eye1 D2 does so well on a wide gamut, as is

look at the crazy results TFT gets on wide gamut monitors using standard software and a Lacie badged EyeOneD2

maybe it works ok on your NEC because the monitor is using the internal factory measurements for the primaries as a basis step???
Logged
Steve Weldon
Sr. Member
****
Offline Offline

Posts: 1445



WWW
« Reply #75 on: June 27, 2010, 11:13:08 AM »
ReplyReply

Quote from: LarryBaum
I don't think an eye1 D2 does so well on a wide gamut, as is

look at the crazy results TFT gets on wide gamut monitors using standard software and a Lacie badged EyeOneD2

maybe it works ok on your NEC because the monitor is using the internal factory measurements for the primaries as a basis step???
The 1id2 works fine on my NEC's (2) because the SVII software includes matching profiles that allow the 1id2 to work fine within tolerances) using the NEC(s) monitors and 1id2 puck.. Without these matching profiles they likely would work they same as they work on other wide gamut displays.. like crap..

Not only that.. I have (2) LCD2690uxi2's side by side in a dual display stand.. and they're matched perfectly.  So.. if they are wrong.. they're wrong the same amount.. and I an easily make up this difference in the rest of my color profile.  People need to remember the monitor profile is only the beginning..
Logged

----------------------------------------------
http://www.BangkokImages.com
Pages: « 1 2 3 [4]   Top of Page
Print
Jump to:  

Ad
Ad
Ad