Removing color cast from your icc profile

Did a little reading up on the net, and came across this very simple and direct

tact for removing color cast from a generated icc profile:

For example, if your results with the profile show a green cast: open the

target scan in Photoshop, start a curve dialogue, and do a simple pull of the

midpoint of the green curve, to make it *brighter*, say from 118 to 138.

Save as a copy, and create a new icc profile with this file. Try it, observe

the result, repeat the process revising the curve adjustment as need.

So in a nutshell, whatever the cast your icc profile is imparting, move the

color of your target scan *in* that direction, so that your icc profile

creating software will move the color of the profile it creates in the opposite

direction.

Mendel, I assume you could use this procedure on any print that used that icc profile and then apply the correction to the image used to generate the profile. Is this correct?

I like this procedure much better than using the profile editor since I am more comfortable with Photoshop and I have more options available to modify the color - curves, levels, hue, etc.

Hi Robert. I really can't comment on your first paragraph, I've done zero printing and have no experience. My immediate issue is scanning Kodachrome, and the off-color results I'm getting.

I've obtained a Kodachrome target, and profiled my scanner with it, through Vuescan. I then apply this profile to my output, through Vuescan.

I had (for example) one scan, a beach picture, with a lot of grey sand in the picture. The histogram had 3 definite peaks, roughly in the middle, due to this expanse of sand. My output with this image had an obvious green cast. Seeing the green peak of the histogram well off to the right of the red and blue, I was looking for a fix. Googling:

"kodachrome scan cast"

The second result was this article:

www.hutchcolor.com/PDF/Kodachrome_profiles.pdf

Which outlines the full procedure behind what I described in my first posting. This seems quite involved, but the first thing I took away from it was that you can modify the icc profile you've made (through whatever software), by modifying the target the software's looking at.

Trying this last night, I believe I was being a bit naive about how much I could accomplished by my simple approach: my results were getting more neutral (I think), but suddenly I had white skies blowing out magenta. It could very well be I've just pushed it too far. My procedure was very crude, compared to the pdf article.

Anyway, I think the linked pdf worth a read. It describes a very careful procedure, and strongly suggests you use a particular light table in the process. The other google results look interesting as well.

Mendel, thanks for the input. I did not realize you were talking about a scanner profile.

I have a Nikon 9000 scanner, but have only scanned a few Kodachrome slides to date. I am using Nikon Scan 4.0.2 and did not get any blue cast with the few I tried. I am scanning color negatives now and will scan my positives after I complete the color negatives. I have only scanned about 3000 images and have a lot more to complete this project.