Nikon Scan 4 Color Management Confusion

[ken_b2]

I recently purchased a Nikon Coolscan V ED to scan 35mm slides for archival

purposes. I only want to do this once (thousands of slides), so Iﾒd like to get

it as ﾓrightﾔ as possible. I use Photoshop Elements (PSE), not Photoshop, as my

editor. But ideally, I'd prefer to archive now, and touchup as needed later.

 

Iﾒm scanning at 4000 dpi, 14-bits per color, but Iﾒm not sure which color space

to use. I started out using the widest, Wide Gamut RGB, but Iﾒm now confused.

 

Wide Gamut RGB looks good when viewed within Nikon Scan, but when the saved file

is displayed outside of Nikon Scan (e.g., PSE), it looks quite a bit desaturated.

 

Iﾒve since played with most of the other available (2.2 gamma) color space

profiles, which has led to additional confusion. Adobe RGB, CIE RGB, and Wide

Gamut RGB all show up as "untagged" in PSE. Oddly, the only one other than sRGB

that seems to include a tagged profile is "Scanner RGB", even though the manual

explicitly states "[when Scanner RGB profile is used] an ICC profile [is not]

included with the image when it is opened in the host application". Instead,

sRGB IEC61966-2.1 is embedded in the file. I also tried the sRGB profile, which

looks fine, but it doesnﾒt seem the best choice for archiving slides that I

intend to actually discard once completed.

 

I tried Vuescan, but the demo version saves only in sRGB. Iﾒd definitely prefer

not to spend even more money, especially when it seems like Nikon Scan can get

it done, except for this issue.

 

If anyone can answer one or more of the following questions, please jump in!

 

1. Which profile is "best" for archival purposes using Nikon Scan, given what

Iﾒve seen?

 

2. Why does "Wide Gamut RGB" (as well as a few others) look great in Nikon

Scan, but desaturated when saved and viewed elsewhere?

 

3. When using "Scanner RGB" profile, is it really embedding sRGB? What good is

that, then, over just using sRGB to begin with?

 

4a. Is my best choice something like CIE or Adobe, and do not worry about the

fact that PSE shows it as "untagged"?

 

4b. Secondary to archiving, Iﾒd like to get some of these printed. Does CIE

convert well enough to sRGB (which is what I would have to do in order to get it

printed at most places)? First Iﾒd have to convert to 8-bit, and then save it

with sRGB embedded. I believe when I do that, PSE converts the color space to sRGB.

 

Sorry for all the questions, but after a couple of days of reading, fooling

around, and not really getting anywhere, I thought it was time to ask for help!

 

Ken

[Ed_Ingold]

Adobe RGB is probably the best color space for you at this time. sRGB has a smaller gamut, and will clip colors whereas Adobe RGB is a little larger than your monitor and home printers. The wider the gamut, the larger the steps between levels, which can lead to posterization in facial tones. This is not a problem with Adobe RGB.

 

Images in Adobe RGB or Wide Gamut RGB will look faded when viewed in applications which are not color managed. All will look pretty much the same in Photoshop, which is color managed. sRGB is intended for viewing in non-color-managed applications, like Microsoft Internet Explorer, and by default in most minilabs. It is simple enough to make a copy and convert to 8-bits and sRGB for these applications.

 

For a comprehensive reference, see "Color Management for Photographers" by Bruce Fraser. For a more general outlook, see "Adobe Photoshop CS2 for Photographers" by Martin Evening.

[bobmichaels]

Edward gave you all the right answers. Heed his advice.

[ken_b2]

Edward, thank you for your response. I believe I will take your advice, as I�m sure you know and understand much more about this than I.

 

After I posted my question, I ran some tests, and did some more reading. I think part of my problem stems from the fact that PSE uses only 2 color spaces � sRGB and Adobe RGB. So it is color managed, but on a somewhat restricted basis. I�m sure the large part of my problem, however, is ignorance. I will check out the references you recommended � thank you.

 

Comparing sRGB & aRGB, I find that sRGB is more pleasing to my eye. That�s too bad, as I�ll have to work some to get the aRGB images to look �better�. There�s the increased saturation, which is easy, and not necessarily a good thing. But there�s also a color difference, especially in blue. There seems to be more red in the aRGB version, and it doesn�t look as �real� to me (nor does it match the original slide as well). Still, for future use, I suspect I am better off heeding your advice.

 

Anyway, again, thanks for your help.

 

Ken

[frank_oddsocks]

<i>Why does "Wide Gamut RGB" (as well as a few others) look great in Nikon Scan, but

desaturated when saved and viewed elsewhere?</i>

<p>

You don't say whether you're on a Mac or a PC. I'm on a Mac and found Nikon Scan

mishandled the colour profiles. I solved the problem by buying VueScan and pointing it at

the .ICC file that came with the scanner. I have posted this information here before on <a

href="http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00FGsO">this

thread</a> (search for "magenta" in your browser). The Nikon software (i) does not

properly embed the profile it used (ii) when asked to output "Adobe RGB" does the

conversion wrongly, and (iii) when asked to use "Scanner RGB" makes what looks like a

correct conversion to sRGB but wedges the machine doing it.

[ken_b2]

Hi, Frank,

 

I just read the thread describing the problems you had. I'm on a Windows system, and I'm hoping perhaps Nikon Scan works better on it. At least, it seems to.

 

It turns out that the vast majority of problems I encountered were due to my not understanding Photoshop Elements' color management, not with Nikon Scan. The purple shift and the loss of saturation were both due my mis-understanding and mis-use of PSE 3.

 

I've since found that if I work in "Full Color Management" (which uses the adobe RGB workspace), I can convert an aRGB to sRGB with no color shift or loss of saturation. Also, comparing an aRGB converted image to an image with an sRGB profile embedded by Nikon Scan shows only a very small difference in contrast between the two (the aRGB image seems to have deeper color in the shadow areas).

 

That thread you linked to has an excellent article link in it that explains just how confusing PSE 3's color management is. I knew most of what's in the article -- after a lot of digging -- but it's nice to see it spelled out so clearly.

 

Thanks for your response, Frank. Hopefully you've had good success with VueScan. And hopefully I will find I can use Nikon Scan plus PSE 3 without further surprises.

 

Ken

[Tim_Lookingbill]

PSE does not have the capability to convert to another space.

What full color management does is assign the AdobeRGB

working space to any opened UNTAGGED image regardless

what space it is WRITTEN if the capture device doesn't embed

the profile space it wrote the image to.

 

If you know for sure your scanner is writing your images to the

AdobeRGB space then you shouldn't see a preview change

when opened in PSE set to FCM.

 

Limited Color Management assigns/previews thru sRGB to

untagged images. If your images are tagged and PSE sees that

it is, it should honor and preview using the embedded/tagged

profile. Not sure on this. I use PS 7.

 

The limit of PSE is that you can't convert AdobeRGB written

images to the sRGB space for output to a minilab. This is why if

you plan on outputting to these types of outside printers, you

should have your scanner WRITE to the sRGB space from the

get go or purchase full version of PS. Otherwise your AdobeRGB

minilab prints will look dull as if you assigned sRGB to them.

 

If you print to your own printer, there's a way to use the printer

driver and/or settings within PSE to print AdobeRGB written

images to your printer without converting.

 

See? Clear as mud. Why does Adobe do this to us?

[ken_b2]

Tim, the good news for PSE users is, Adobe saw fit to allow for easy conversions between aRGB & sRGB in PSE 4. Or so Iﾒve read; I am still using PSE 3.

 

The color shift that I encountered occurred using "Limited Color Management", which uses sRGB as a working space. What PSE 3 does in this case is it gets rid of any color space profile data in an image if it isnﾒt sRGB, and marks it as "untagged". (That's "limited" all right!) So it was showing me an image created with an aRGB color space profile using sRGB ﾖ no wonder it didn't look very good.

 

Even in PSE 3, there is a way to convert aRGB images to sRGB. It's a bit of a trick, but it does work. In "Full Color Management" mode (which uses an aRGB working space profile), if you copy a file into an existing file, it converts the data to that fileﾒs color working space. So, I simply copy the aRGB image I want to convert and paste it into an sRGB image I already have loaded. Voila! Flatten and Iﾒm done. The aRGB image is actually converted to the sRGB space. (Of course, I've got to be careful when saving the file to be sure the 'save with profile ...' box is unchecked, otherwise it will tag the profile as aRGB after all!)

 

There are several other nuances to all this. Adobe indeed created quite a mess, in their attempt to make things "simple"! Ha!

 

I really can't justify the price of a full-blown Photoshop, so I'm hoping that when PSE 5 finally comes out, they may have made color management even more full-blown. But Iﾒm not holding my breath.

[Tim_Lookingbill]

Ken,

 

Thanks for the insight on PSE. I wasn't aware of the past trick for

converting to sRGB. The stripping of profiles and forced

assigning of sRGB with Limited CM is something I forgot about.

 

The main issue is knowing what space your capture device

writes the image to if they happen to embed the profile in such a

way PSE or PS can't read it or have access to it or if they embed it

at all. Preview mismatches in these types of scenarios create

doubt in both the capture devices' software and the image

editors' making things even more complicated with such a

limited app as PSE.

[don_e]

PSE4 can strip the profile or convert to sRGB or aRGB. If you set up preferences to let the user determine the profile when opening a file, it will preserve the profile (for me that's prophoto). But within the app, you are limited to s and a RGB (or no) conversion.

 

--

 

Don E

[howard_owen]

Note: The following applies to using NS4 with a Coolscan 4000. I can't be certain that the newer scanners exhibit the same issues but my bet is that they do.

 

Nikon's color management is uniformly awful and should be avoided, period. The advice to use Adobe RGB would be sound enough if it were implemented properly in NikonScan 4, but it is not. Nor was it any good in NS3. If you will examine slides scanned in this colorspace, you will find moderate to severe posterization in areas that should contain shadow detail. This is all the reason you need to discard it, especially if you want to scan your slides only once.

 

Nikon's color management should either be turned off, or the profile "Scanner RGB" should be used if color management is desired to be left on. FYI, in the drop down menu in the NS color management dialog, the color profiles are listed in *ascending order* as to gamut width: sRGB, the narrowest, tops the list, Scanner RGB is on the opposite end. According to my contact with Nikon, the fact the images come into Photoshop apparently tagged as sRGB (when Scanner RGB has been selected as the colorspace) is due to Windows mishandling the information. This may or may not be true as Nikon software has its share of issues. However Nikon assures me that the files can be regarded as untagged.

 

My procedure is to discard the profile when opening in Photoshop and then assigning a profile of my choosing (normally Adobe RGB or Holmes Ektaspace). Some of my associates nearly faint dead away at this notion, claiming that I should be *converting* the profile instead. Try both and see which you prefer.

 

Good luck.

[ken_b2]

Ah, just when I thought I had it all figured out, and it was safe to begin scanning!

 

Thanks, Howard, for raising this concern. Clearly it IS a big concern. I checked one of the test slides I did, which has a great deal of shadow s. Some were so dark I had to apply a 50% lightening to see them. But I was not able to detect any posterization. I compared scans made with Nikon Scan against both SilverFast & VueScan. Ignoring the large differences in colors, contrast, etc., I could not see any differences in missing details.

 

My concern is that perhaps I don't know what I'm looking for, since I'm not seeing it. Though I have seen posterization before in other (extreme) situations, so I know what it looks like when it's obvious.

 

Using "Scanner RGB" results in a very washed out looking scan. I can fairly easily correct this in PSE, but it would mean that EVERY scan would need to be corrected. (Though alternatively I could probably find a way to "fix it" in Nikon Scan, too, since "Scanner RGB" does allow you to fiddle around with color correction, unlike what happens with disabling color management altogether. Perhaps I could fiddle with that. If I can find a setting that works, then it might be a way to go. But I'm not going to PSE every scan.)

 

Thanks for the info about sRGB showing up in "Scanner RGB" files. I'm not sure I trust the answer, either, but I suppose it's possible. :)

 

Unfortunately, PSE 3 doesn't allow me to discard the profile upon opening it, nor does it let me save it with a profile other than sRGB.

 

Looks like I've got more research ahead before I can begin the huge job.

 

Thanks,

 

Ken

[howard_owen]

Ken,

 

The posterization is obvious when using Nikon's Adobe RGB profile. If you aren't seeing it, it probably isn't there. It's possible that Nikon addressed the problem with the most recent scanners. You probably know that many of Nikon's profiles are proprietary (the ones with an underscore in their name), meaning that you cannot apply them in other software outside of NS. The reason for this is probably two-fold. One, the profiles do not conform to the icc standard. Two, Nikon does something to the image data in the scanner's firmware before it's passed on to Photoshop which makes its profiles unsuitable for other uses. This latter point was seemingly verified by a member of another forum community a few years ago when the whole Nikon CMS can of worms was a little fresher.

 

The bottom line here is that if you are able to produce satisfactory scans with your current work flow, there's no need to change it. I just wanted to caution you before you scanned several dozen, or several hundred, slides (as I did), and then found a serious flaw.

 

PS: If you want to sift through some posts on the topic, you can go over to

http://groups.google.com/group/comp.periphs.scanners?hl=en&lr=&ie=UTF-8&safe=off

 

and run "Nikon color management" through the search engine, or just google it. When I did that, I discovered that NS adds the sRGB tag to files scanned on a Mac so apparently it's not a Windows thing. To quote, "Photoshop... reads info from color profile and EXIF (Exchangeable Image File Format http://www.exif.org/) data. When color management on Nikon scanner is turned off NikonScan still specifies sRGB in EXIF data and Photoshop reads this data and interprets it as if image had sRGB profile embedded." I assume this applies when using Scanner RGB as well.

[ken_b2]

Wow, Howard, what a mess. I did some reading on the newsgroups to look at the problems people were having. But it does look like 4.0.2 fixed at least some of the more obvious ones. Besides the posterization I'm not seeing, I also notice that NikonScan comes with many non-underscored profiles. Though I'm not sure that necessarily means they're all non-propietary. But the aRGB profile (NKAdobe.icm), which is named "Nikon Adobe RGB 4.0.0.3000", seems to be the same as the standard "Adobe RGB (1998)". I used the rather neat online VRML tool called "ICCView" at this site: http://iccview.computertobi.de/ From what I read, it seemed these non-underscored profiles didn't used to be included with NS.

 

Anyway, I can certainly understand why SilverFast & VueScan became so popular with Nikon users. :-)

 

Thanks again,

 

Ken