Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: Nill Toulme on May 10, 2006, 09:46:56 pm
-
I've heard it both ways. Which is better (and why)?
I'm using an NEC 2090uxi and Eye One Display 2.
Thanks,
Nill
~~
www.toulme.net
-
I think common wisdom is native, but I go with 6500 because I want all my monitors, regardless of type or brand, to match.
-
Native White Point is generally better for LCDs because you're not adjusting the white point via the graphic card but rather using Photoshop's Display Using Monitor Compensation architecture which does this with higher bit precision. Another "issue" is 6500K is kind of an ambiguous value (there are many colors that correlate to this value).
-
Native White Point is generally better for LCDs because you're not adjusting the white point via the graphic card but rather using Photoshop's Display Using Monitor Compensation architecture which does this with higher bit precision. Another "issue" is 6500K is kind of an ambiguous value (there are many colors that correlate to this value).
[a href=\"index.php?act=findpost&pid=65040\"][{POST_SNAPBACK}][/a]
Thanks Andrew. Does this remain true using other color managed apps? I do most of my color work in the RAW conversion with C1.
Gandalf — that makes sense, but I'm only doing color work on the one monitor so I don't have the monitor-matching issue to worry about.
Nill
~~
www.toulme.net
-
Thanks Andrew. Does this remain true using other color managed apps? I do most of my color work in the RAW conversion with C1.
[a href=\"index.php?act=findpost&pid=65041\"][{POST_SNAPBACK}][/a]
It should.
-
OK thanks again.
Nill
~~
www.toulme.net
-
What's the point of buying an LCD with internal 12 bit LUTs? Isn't it to be able to losslesly calibrate to any target you choose? Am I missing something?
NEC 2090uxi has 12 bit internal LUTs.
-
What's the point of buying an LCD with internal 12 bit LUTs?
[a href=\"index.php?act=findpost&pid=65047\"][{POST_SNAPBACK}][/a]
The internal LUT is used for gamma adjustment and neutrality. I'm sure the backlight has a specific whitepoint, but don't know whether the controls on the NEC allow you to bypass the LUT to utilize it. Maybe it doesn't matter.
-
Well how are they used for gamma adjustments when there is no DDC connection? And "neutrality" IS the white point. It really looks like in this particular case having a 12 bit LUT monitor has no benefits whatsoever. It looks like everything is done through the videocard LUTs... Unless the onscreen controls affect the internal LUTs... (?) Sounds unlikely...
If the adjustments are done through the videocard then Native gamma and Native whitepoint are your best bet of course - meaning just go ahead and measure the factory default output. Otherwise visualise applying a Curve to an 8 bit PS file and then looking at the Levels histogram - you'll see gaps, values that are no longer there. That's what happens when you adjust gamma and white point through videocard LUT.
-
Well how are they used for gamma adjustments when there is no DDC connection?
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=65053\")
The manual for the 2090 is here (warning 7MB file):
[a href=\"http://www.nec-display-solutions.com/coremedia/download/164524/2090-UsersGuide-english.pdf]http://www.nec-display-solutions.com/corem...ide-english.pdf[/url]
According to this you can adjust both the whitepoint and gamma (from 0.5 to 4.0). It's under Advanced menu, tags 5 & 6.
The 2090 also supports SpectraView II for DDC hardware calibration.
-
Thanks. It appears that to take advantage of DDC you need to use this setting for Color Control:
"PROGRAMMABLE: The colour tone that was set up with the downloaded application software is reflected."
Which probably means the Spectraview software.
I've seen on another forum a post claiming that only Spectraview models (the manual you linked to is for a Spectraview model) have this feature and it will not work on a non-Spectraview monitor.
It still does not mean RGB sliders don't affect internal LUTs though... Or perhaps the Naviset software bypasses the videocard LUTs... But you definitely have to be able to get more value out of this monitor than just calibrating to all Native...
-
I've seen on another forum a post claiming that only Spectraview models (the manual you linked to is for a Spectraview model) have this feature and it will not work on a non-Spectraview monitor.
[a href=\"index.php?act=findpost&pid=65057\"][{POST_SNAPBACK}][/a]
The manual is for both. I assume the difference is that the SpectraView model comes bundled with the software.
-
This is the post I was referring to:
http://forums.dpreview.com/forums/read.asp...essage=18371820 (http://forums.dpreview.com/forums/read.asp?forum=1004&message=18371820)
I just feel that with all the technology incorporated in this monitor you gotta be able to have more flexibility than Native. But than again maybe not... They say it's firmware...
[edit] If it's firmware there has be some workaround... Somehow it has to work with DDC.
-
This is the post I was referring to:
http://forums.dpreview.com/forums/read.asp...essage=18371820 (http://forums.dpreview.com/forums/read.asp?forum=1004&message=18371820)
[a href=\"index.php?act=findpost&pid=65059\"][{POST_SNAPBACK}][/a]
The link didn't work for me, but I found the message anyway:
"The SpectraView Profiler software can basically be used with the standard
MultiSync LCD2090UXi, however this will be software calibration (grafic
board side) only.
For hardware calibration you need to use the SpectraView 2090.
This is one of the differences of these two models. Beside this the
SpectraView models have i.e. more narrow specifications for uniformity and
pixel failures."
It looks like NEC wants to retain control over the hardware calibration and keep the profits in-house. Also maybe why support for NEC models was pulled from basICColor's Display. Sigh. My understanding is that the Custom functions are still supported by the base model ... maybe Nill can confirm this.
-
Sorry. Yes, this is the post. Sigh...
-
Thanks all. The Spectraview question will require some more digging. I don't think the latest version with 90-series support is out yet. It was due out this month. I have a question in to tech support about it.
FWIW, the press release (http://www.necdisplay.com/press/PressDetail.cfm?document_id=2247) for the 90-series says:
The new “i” series NEC Professional 90-Series features include:
• Standard four-year limited warranty
• SpectraViewII, which delivers an advanced and easy-to-use calibration and profiling solution for color-critical applications
Nill
~~
www.toulme.net
-
And "neutrality" IS the white point.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=65053\")
Not really, it's the color of white at the target luminance.
As for bits, read this from Karl Lang:
[a href=\"http://luminous-landscape.com/forum/index.php?showtopic=9613&hl=]http://luminous-landscape.com/forum/index....wtopic=9613&hl=[/url]
-
Not really, it's the color of white at the target luminance.
Yes, but from what I understand it determines what "neutral" is.
-
Yes, but from what I understand it determines what "neutral" is.
[a href=\"index.php?act=findpost&pid=65133\"][{POST_SNAPBACK}][/a]
Neutral can run anywhere from black to white...
-
Neutral can run anywhere from black to white...
[a href=\"index.php?act=findpost&pid=65142\"][{POST_SNAPBACK}][/a]
I admit I don't understand this statement.
-
I admit I don't understand this statement.
[a href=\"index.php?act=findpost&pid=65152\"][{POST_SNAPBACK}][/a]
Try this:
Create a new document in Photoshop. The size isn’t critical but try to make
the horizontal about 1200 pixels. You want to view as much of
this document at 100% zoom ratio in Photoshop filling most of
the display you are testing. The color space of the document
isn’t important.
2. Choose Edit-Fill and when the Fill dialog appears select Use:
Black, BlendingMode: Normal, Opacity 100%.
3. Choose Image-Mode-Assign Profile in Photoshop CS or
Edit-Assign Profile in CS2. From the list of profiles you need
to select your display profile. What this does is ensure that
Photoshop will simply send the numbers in this document
directly to the screen. This will aid in evaluating the display
profile.
4. You will need to select the Rectangle Marquee from the
Photoshop toolbar or simply hold down the M key. Make sure
that in the option bar the Feather is set to zero and that the
Style: pop-up menu is set to Normal.
5. You will need to make a rectangle selection in this document. It
can be just about any size but try and make it so it fills about 50 percent
of the document in the center. You will need to be able to see inside and outside
of this marquee so don’t make it too large or too tiny.
6. Hide the marching ants. This can be done by selecting View-
Show-None or by holding down the Command/Control H
key.
7. At this point, You need to fill the entire display with this black
document so we need to go into full screen mode. The easiest
way to do this is to press the F key twice. Hide the tool bar and palettes. By holding down the Tab key,
8. At this point, you should have a screen filled with only black!
You we need to call up the Curves command. Since the
menus are hidden, you’ll need to use the Key command,
which is Command/Control M. The curves dialog should
appear. Move it over to a corner since you will want to see the
effect of the hidden rectangle marquee.
9. At this point, you will move the black curve point up one
numeric value at a time. To do this, simply press the up arrow
key on your keyboard once. The output value will change from
zero to 1. If instead the values pops up to 2, enter 1 into this
output field and all subsequent pressing on the up arrow will
ensure only one value is added. Keep pressing this up arrow
key until you can just see the separation from the rectangle
marquee. That is, you want to just barely see the rectangle. You
may need to press the up arrow key several times. Ideally, you
want to see separation between a value of 0 and 1. I can see
this on the Artisan.
Keep moving the up arrow slowly once you see the separation.
The next thing you want to do is examine how neutral each
step is. Not only should each step show the same density
difference from the last, but also each progressively gray step
should be neutral. It is not uncommon to see a gray square go
from neutral, to slightly green, to slightly magenta, and back but that's NOT ideal.
-
Some ready-made files people may find helpful for determining display/print neutrality and black separation:
http://www.macquarieeditions.com.au/misc/RGBGrays.tif (http://www.macquarieeditions.com.au/misc/RGBGrays.tif)
http://www.macquarieeditions.com.au/misc/LABGrays.tif (http://www.macquarieeditions.com.au/misc/LABGrays.tif)
http://www.macquarieeditions.com.au/misc/Blacks.tif (http://www.macquarieeditions.com.au/misc/Blacks.tif)
For the RGB files, any profile (display or working space) can be assigned. The first two I think I found on Bill Atkinson's site.
-
Thank you Andrew. I feel like this theoretical discussion kind of hijacks the thread even though I'm extremely interested in it. The original question of if it's possible to losslessly calibrate a 12 bit LUTs NEC 2090uxi to any white point and gamma (and how to do it) is still open.
Since I am really interested in this kind of tests I went ahead and ran it (it's easier with two monitors). Obviously a Dell LCD is not an Artisan and Spyder2 is not Coloreyes+Xrite. So yes I only see differences starting from 7 and yes I see very perceptible differences in tint going through the shades of grey.
However... As far as the controls I have over the calibration the White Point Target is the only control I have that affects neutrality. So even if it works imperfectly there's not much I can do to change that.
One more thing I am not clear about - why look at the test images in monitor profile? It essentially simulates non-colormanaged applications, correct?
I don't know what to look for in the last two of Stephen's images - I know my greyscale ramps are not perfect but tolerable.
-
One more thing I am not clear about - why look at the test images in monitor profile? It essentially simulates non-colormanaged applications, correct?
[a href=\"index.php?act=findpost&pid=65160\"][{POST_SNAPBACK}][/a]
Andrew is more capable of answering this but, as I understand it, the intention is to take the monitor profile out of the picture so you're looking at separation and neutrality of the display by itself. The less work the LUT in the display card has to do, the better.
As for the files I referenced, you can use these in a number of ways. The LABGrays (and RGBGrays) file is useful for assessing monitor profiles. Better profiling packages will generate profiles that are more neutral (look cleaner) and smoother (less banding) but this is just one criterion for evaluation. The Blacks file can show you if and where your shadows are blocked, both on display and in the print. Note that the distribution will change with the gamma of the assigned working space. Have a play around with a number of monitors and profiles to get a feel for what these are telling you.
-
I wrote the post in dpreview quoting Alexander Klein. I got the same reply from the basICColor team, so I assume it's correct. I therefore have been wondering about the purpose of the 12bit LUT as well for the models that do not have the proper firmware for DDC.
I'm personally waiting for the SpectraView 2090 model to be released so I can compare prices to the 2090uxi, but it will probably be too expensive for me.
Since Nill has only said good things about the regular version I'm not too worried though
-
Marc, this is interesting. I am convinced there's gotta be more value in this monitor than a typical LCD. It's surprising how little information is available.
Stephen, thanks for the explanation. I think Photoshop does not affect the LUTs though. I'll experiment with your files some more.
-
Marc, this is interesting. I am convinced there's gotta be more value in this monitor than a typical LCD. It's surprising how little information is available.
Stephen, thanks for the explanation. I think Photoshop does not affect the LUTs though. I'll experiment with your files some more.
[a href=\"index.php?act=findpost&pid=65458\"][{POST_SNAPBACK}][/a]
The confusion about the 1990SXi/2090UXi/2190UXi being hardware calibratable over DDC with the Spectraview stems from different NEC policies in Europe vs US:
- The US Spectraview II software is completely different from the European Spectraview Profiler 4 software. The latter is an OEM version of basICColor, while the former is developped by NEC US.
- European and US versions of the 9 series displays have different firmware tags, so that hardware calibration software can discern them.
- In Europe, the Spectraview Profiler software is only sold in the Spectraview packages. The displays in the Spectraview packages has a different firmware tag then the non-Spectraview ones. The Spectraview Profiler refuses to hardware calibrate the displays that don't have the European Spectraview tag in their firmware (so it also refuses the US models). This is all due to recent change in the NEC Europe policy. The company making Spectraview Profiler is no longer allowed to release a version of its own software, basICColor, that is capable of hardware calibrating the non-Spectraview 9 series (and not even the old 8 series).
- In US, the Spectraview II software can be bought separately and is supposed to hardware calibrate the non-spectraview 9 series displays (= the ones not sold in a Spectraview package. However, the latest version is not released yet, so it's not clear if that policy will be maintained.
-
- In US, the Spectraview II software can be bought separately and is supposed to hardware calibrate the non-spectraview 9 series displays (= the ones not sold in a Spectraview package. However, the latest version is not released yet, so it's not clear if that policy will be maintained.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=65489\")
Interesting and helpful, thanks. Based on my recent exchanges with US NEC support, it seems likely that the policy you describe will be maintained. They're still predicting release of the new version of Spectraview II later this month.
Nill
~~
[a href=\"http://www.toulme.net]www.toulme.net[/url]
-
Great, so people in the US get to DDC calibrate with the regular monitors while us in Europe have to pay through the nose for the SpectraView version of the monitors?
That sucks.
-
That sucks.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=65568\")
Well maybe, maybe not. As I recall the SV versions include both the software and the hood. Those two things add up to an additional 320 clams or so in the US, so unless the EU price delta between the SV and non-SV versions is a lot more than that, it's a wash — other than the fact that there's no "upgrade" path for non-SV owners I suppose.
I wonder what keeps a European customer from using the US software?
Nill
~~
[a href=\"http://www.toulme.net]www.toulme.net[/url]
-
I can make myself a cheap hood, and the price difference between the two monitors here is 500 usd while I could pick up the software for less than 200.. since I'm already stretching the budget it's not so fun for me.
I wonder what keeps a European customer from using the US software?
Well, since the firmware in the European 2090uxi is different from what I understood there's no point getting the US software since it won't DDC with the altered firmware.