settings to match color of scan to original slide

[jenlrile]

I've got a Nikon 8000, Silverfast 5.5.2r20 with IT8 module (using Photoshop plugin), Photoshop 7.0, and Windows XP SP1. I've got a 5000K light table and NEC MultiSync FP1375 monitors calibrated with NEC's calibration software and colorimeters.

I'm having some trouble getting the scanner, Silverfast settings, and profile set up so that scans match the color of positive slides when viewed on the 5000K light table. In other versions of Silverfast with other scanners (e.g. Nikon 2000, a few flatbeds with transparency adapters), the following settings have worked well:

Auto tab:
Gamma: 2.2
CMS tab:
Scanner -> Internal: ICM
Internal -> Monitor: ICM
Internal -> Output: <RGB>
Scanner (Transparency): SF_T (whatever scanner we're using) [the profile created with the IT8 module]
Internal: sRGB
Rendering Intent: Perceptual
Embed sRGB ICC profile

ICE is off, and histogram and gradation dialogs are reset both for the calibration procedure and the scan. The color simply does not match, even when scanning the target immediately after creating the profile with the IT8 module. The scans have a strong dark blue cast.

I've tried wiping and reinstalling Silverfast, deleting the old scanner profiles from the system, running the IT8 module with and without autoexposure on, scanning with and without autoexposure on, changing the Lightness settings on the Special tab, and probably a few other things I can't remember right now. Color still doesn't match. It also doesn't match when the Scanner -> Internal setting is Calibration. I've run the IT8 module with two separate 35mm Ektachrome IT8 targets and associated data files, both less than 2 years old and stored without light exposure.

The same color shift seems to be happening with all film types we try - most recently some 1960's-era Kodachrome (I know Kodachrome is very blue, but the scans still should match what I see on the light table!), but also modern Ektachrome and various other unidentified positive films we get from time to time. Again, even scanning the IT8 target itself does not yield acceptable results.

I know that calibration isn't perfect, but the results I'm getting here are WAY off from all of our other equipment. So are there some new settings with the Nikon 8000 that I'm mis-interpreting? Auto Contrast is unchecked. Limit Gamma Slope is unchecked. What else can I try?

Thanks for your help. -Jenn

[LSI_Support]

Don't use the sRGB color space, use Adobe RGB 1998 instead.

[jenlrile]

Thanks for the suggestion. I tried creating the profile and doing subsequent scans using the Adobe RGB color space, but there was not a significant increase in color accuracy.

Does anyone else have suggestions for fixing this problem?

Thanks -jenn

[degrub]

You might cross check your CMS settings with those that Ian Lyons gives in his tutorial for CMS under Silverfast.. www.computer-darkroom.com

Frank

[jenlrile]

Thanks for the suggestion, degrub. I've consulted Ian Lyons tutorial on CMS settings. He does a great job of reminding us that settings are dependent on what you're trying to do.

I've got the IT8 module for the Nikon 8000 (and all the other scanners we use), and we've always used settings similar to what Ian L. shows in his CMS Example 5. (See my first post for exact settings.) They also match what is suggested in the Silverfast documentation for a color-managed workflow. The one difference is the Internal -> Monitor setting. We've got calibrated monitors using Windows ICM for management, so we set this to ICM, rather than Ian's suggested Automatic. As I stated before, I've tried the Scanner -> Internal setting on both "calibration" and ICM, using the profile created by SF's IT8 module, but I'm still not matching color.

I'm concerned that there are some settings in the Nikon 8000 version of Silverfast that could be affecting color reproduction that I have set wrong, but don't know it. I've got the "Lightness" dialog set all to the default 0's. What about autoexposure? This obviously will affect the color. Does it need to be on or off when creating the profile? When scanning using the profile to match image color to the slide? I'm not getting good results either way. Other settings I need to be considering and/or testing?

Thanks,

-jenn

[milstone]

Are you sure the iT8 module is working at all? You could try scanning with and without iT8 to be sure that it modifies the color calibration, even if the end result is still not correct.

[ilyons]

The one difference is the Internal -> Monitor setting. We've got calibrated monitors using Windows ICM for management, so we set this to ICM, rather than Ian's suggested Automatic. As I stated before, I've tried the Scanner -> Internal setting on both "calibration" and ICM, using the profile created by SF's IT8 module, but I'm still not matching color.

You do NOT want to choose ICM - you SHOULD choose Automatic and allow SilverFast CMS and Preview window to properly integrate with Photoshop. By choosing the ICM option you are previewing the image in monitor RGB which isn't anywhere near how the image will appear in Photoshop.

I'm concerned that there are some settings in the Nikon 8000 version of Silverfast that could be affecting color reproduction that I have set wrong, but don't know it. I've got the "Lightness" dialog set all to the default 0's. What about autoexposure?

Auto Exposure and IT8 CM don't really fit well together. The whole idea of IT8 is to profile the scanner in a state that will remain constant. Auto Exposure is as constant as the weather!

This obviously will affect the color. Does it need to be on or off when creating the profile? When scanning using the profile to match image color to the slide? I'm not getting good results either way. Other settings I need to be considering and/or testing?

From your earlier post you appear to scanning some Kodakchrome. The sad truth is that unless you have the appropriate IT8 slide and data file for this film your results with an Ektachrome IT8 slide will make matters worse. The results will be horrid! I've done a Kodachrome profile recently for one of the other forum members - maybe if they see this post they'll share the profile. Even then I hd to make a very substantial edit to the profiles blue channel. Kodachrome really is a horrid film when accurate colour rendering is required from scanning.

[milstone]

I have been using autoexposure and iT8 together with pretty good color accuracy results but variable overall brightness/contrast results. Emphasis on variable. This is easily compensated in Photoshop but has always been perplexing. I even asked about this on this forum.

The comment that autoexposure and iT8 are not entirely compatible makes sense. I will eliminate autoexposure and see if that improves brightness/contrast consistency. It seems this might not affect color accuracy if the color spectrum of the lamp were constant at different illumination intensities. Is this the correct way to think about it? On the other hand, the color spectrum probably varies with intensity. Correct?

I don't remember seeing autoexposure/iT8 compatibility discussed in the manual when I initially set things up. Is this mentioned in the manual? It seems an important point.

[ilyons]

[quote="milstone"It seems this might not affect color accuracy if the color spectrum of the lamp were constant at different illumination intensities. Is this the correct way to think about it? On the other hand, the color spectrum probably varies with intensity. Correct?[/quote]

It's a while now but my memory is of highlight details being retained that were otherwise lost (blasted to outer space in some case), albeit at the expense of neutrality - the last white patch on the IT8 had a mild magenta cast which is consistent with what I've seen on many other scanners . Generally colour rendering is the same with or without Auto Exposure. The final result may lack the bite/punch that you currently get.

I don't remember seeing autoexposure/iT8 compatibility discussed in the manual when I initially set things up. Is this mentioned in the manual? It seems an important point.

It's not dicussed in the manual. I don't think the manual discusses any of Nikons special functions - for that you need to read the Nikon Addendum

[jenlrile]

Hello all;

Thank you for all the comments and advice provided over the weekend.

Milstone, I also suspected that the IT8 module simply wasn't working, as did Silverfast support. But scans with Scanner -> Internal set to <none> are significantly different than with it set to Calibration or ICM. So it's at least *partly* working.

Ian, thanks for the info on autoexposure. Your explanation matched what I suspected was the case. It's good to hear some confirmation of that. Changing the Internal -> Monitor setting to Automatic doesn't solve the problem I'm having. With both the ICM and Automatic settings the image in Silverfast matches color well to the image in Photoshop. I have a question about this setting, though. As I understand it, Photoshop removed its settings for monitor profiles in version 6 and now relies on the OS (or Adobe Gamma Loader if you're using it, which we're not) for information on the monitor color space. Is my understanding correct? If so, how does Silverfast get information from Photoshop v. 6 or 7 on the monitor color space when setting Internal -> Monitor to Automatic?

Ian, you also asked about our scanning Kodachrome. Yes, I did mention that we'd had this problem with Kodachrome, but we are also having it with everything I have available to test, including the Ektachrome IT8 target used for calibration itself. The user you created the Kodachrome profile for did send it to me (thank you!) and I will test it out, but that doesn't seem to be the root of this problem. If it looks like we'll be getting a lot of Kodachrome, we'll test your profile and either use that or buy a Kodachrome target from Kodak and use that for the IT8 calibration.

The more testing I do it seems that my problem is like milstone's - more rooted in brightness than color. But adjusting the lightness settings in Silverfast blows out the highlights almost immediately, and the scans are still WAY darker than the slides on a 5000K light table dimmed way down almost to nothing.

I've contacted Silverfast support and we have some communication going. Hopefully we'll find a solution I can post here. In the meantime, I'm open for more suggestions of things to try!

Thanks,
-Jenn