Gray image has color even when eye dropper says none

[Brian K]

Ok, I have a weird problem. I worked on an image in PS, a gray scale image saved as 16 bit rgb (srgb) and when I went to re-open the image, it had been in the Large File format, PS told me that some information was not readable. Nevertheless the image opened up and looks fine except there's a bit of color in it. A slight shift from green to magenta. However the image reads as absolutely neutral according to the eye dropper. I have changed the image to gray scale, saved and then changed it back to RGB and it still has the color in it. The problem is that it prints rather warm, when it's supposed to be neutral and the slight color shift of green to magenta is subtle but visible in the sky. It's not a printer or paper issue because other prints come out dead neutral using this profile and I've seen it already on 2 different brands of paper.

I am assuming that some information related to the file got corrupted, the file itself seems intact. Anyone ever have this issue?

[Mark Barendt]

Has your monitor and printer been calibrated?

[Ken Lee]

Does the color still appear when the image has been converted to Grayscale mode and saved in that mode ?

Perhaps you should create a new grayscale image, then copy the old image and paste it into the new one, then save it. Make sure the grayscale image has either no profile, or a grayscale profile. Then convert to RGB.

Does the same problem appear in ordinary PSD format ? Or TIFF ? Perhaps converting from one format to another will eventually result in the color channel data being discarded once and for all.

Another option is to perform a Save As and provide a new file name, rather than just Save. This will result in a new file being written from scratch: that's a different process than updating an old file, and may do a better job.

It wouldn't surprise me if Adobe spends much more time QA testing PSD and JPG formats than other less popular formats: I mean, assigning Quality Assurance Engineers. If this scenario has escaped their attention, perhaps it's a pathway rarely treaded.

[Brian K]

Has your monitor and printer been calibrated?

Mark, the monitor is calibrated and other images do not have this issue. Further, the PS eyedropper reads the image as dead neutral. And it's not so much a screen issue, the problem is that on various papers it prints like it's sepia with a little red added. And the profile is dead neutral, (4 printer linearizations, 2200 color patches read for calibration) so what should come out is dead on neutral. But with this one file it dos not.

[Brian K]

Ken it appears when converted to Gray, to cmyk, to LAB, and btw the AB and values in ab are neutral and equal. And it also appears that way as a tif or a PSD. I've even copied a channel from the corrupt file and then pasted it into a new file , in R,g and B channels, which should result in a dead neutral image as all 3 channels are exactly the same. And it still has the issue.

[Mark Barendt]

Mark, the monitor is calibrated and other images do not have this issue. Further, the PS eyedropper reads the image as dead neutral. And it's not so much a screen issue, the problem is that on various papers it prints like it's sepia with a little red added. And the profile is dead neutral, (4 printer linearizations, 2200 color patches read for calibration) so what should come out is dead on neutral. But with this one file it dos not.

The eye dropper is a good indicator of neutral, better even than a calibrated screen.

How about a proofing setting in PS?

[bob carnie]

Brian if you are using a Mac go to utilities and open the digital colour meter and set values to CIE Lab with a small apeture, not all the way otherwise you will be getting point readings.

Open your image in PS and verify the info pallete readings with this meter , they should be the same.

When you say the values in the AB are neutral and equal do you mean 0- a 0-b??? if they are any number above or below 0 but equal you have tone. sorry if this is obvious.

[bob carnie]

Can you post the image?

[Brian K]

Brian if you are using a Mac go to utilities and open the digital colour meter and set values to CIE Lab with a small apeture, not all the way otherwise you will be getting point readings.

Open your image in PS and verify the info pallete readings with this meter , they should be the same.

When you say the values in the AB are neutral and equal do you mean 0- a 0-b??? if they are any number above or below 0 but equal you have tone. sorry if this is obvious.

Bob, I can't work on PS this minute because I'm heading out, but I wanted to post the image. If LAB a or B have a value, and I'm trying to recall if there was a tone in them, and I think all I saw was an even gray, no detail but an equal value for each, can that be repaired or corrected?

I'll try your recommendation as soon as I return. Thanks

[bob carnie]

Brian I am reading 0 values in the A B channel so neutral with no colour values showing as you state.

I do know that when I calibrate 21 steps to neutral on my lambda using colour paper , even though the RGB graphs calibrate out as neutral, my eye perceives a colour tint.
This is the downfall of trying to make BW images on colour paper. One learns to live with the tint or in PS bias the colour to lets say sepia or selenium tone.

The same will happen with an inkjet printer .. your eye will perceive colour even though you are sending over a file as you have posted.
If one wants pure black whites then set your printer to Black White mode rather than colour.

I print silver wet prints off files so I do not do a lot of black white prints on ink or RA4 paper. This is the only possible answer I can give you at this point.
Lots of printers use the piezo multiple black ink sets to keep the work neutral, I have never gone down this road as I prefer the silver prints.
Not saying they are better just what I have evolved my Lab into.

[Mark Barendt]

I think Bob has a great point and I think it's confirmed to a point given that differing papers affect the result.

May be an example of metamerism, too.

Although I have had B&W prints that I really like on RC and from inkjet, when compared to neutral silver prints there has been a difference.

[Brian K]

Guys the problem is that other files using the same profile, printer and paper are dead neutral. The rip software I use is StudioPrint 14 and the profiles are not stock manufacturer ones but custom made ones, using the ColorGPS software included in StudioPrint. The profile building consisted of linearizing the color print out to get the maximum gradation, that is reading 48 patches of each color and then, printing another linearization and fine tuning the gradation, and doing this 4 times. Then printing out and reading 2200 patches of just about every color mix you could imagine. Trust me, on other files, it's dead neutral. The paper is Canson Fiber Platine. It seems a very neutral paper, and other images printed on it appear neutral. But this file is clearly sepia.

Is there any way to edit the color information tagged to the file? Because while the file reads neutral in PS, it does not appear dead neutral on the screen or on a print. I'm thinking that some instruction attached to the file is giving the hardware the wrong information.

I just checked the color using the Mac's digital color tool, set to CIE LAB, I get L 33.962, A- 0.297 and B- -0.391. i some places

[bob carnie]

Set the readings to 8 bit to get your a and b numbers its probably set in 16bit right now.

You should only be getting 0 .. when I read your file I only got 0's in ab

At this point you could adjust colour on the printer to get more neutral if the cast is global, or you could go back into PS and colour correct by looking at the print and ignoring what your eyes are seeing on screen.

Or go back to original scan or process and restart.
Wish I could help more.


Guys the problem is that other files using the same profile, printer and paper are dead neutral. The rip software I use is StudioPrint 14 and the profiles are not stock manufacturer ones but custom made ones, using the ColorGPS software included in StudioPrint. The profile building consisted of linearizing the color print out to get the maximum gradation, that is reading 48 patches of each color and then, printing another linearization and fine tuning the gradation, and doing this 4 times. Then printing out and reading 2200 patches of just about every color mix you could imagine. Trust me, on other files, it's dead neutral. The paper is Canson Fiber Platine. It seems a very neutral paper, and other images printed on it appear neutral. But this file is clearly sepia.

Is there any way to edit the color information tagged to the file? Because while the file reads neutral in PS, it does not appear dead neutral on the screen or on a print. I'm thinking that some instruction attached to the file is giving the hardware the wrong information.

I just checked the color using the Mac's digital color tool, set to CIE LAB, I get L 33.962, A- 0.297 and B- -0.391. i some places

[Ken Lee]

It sounds like a bug in Photoshop. I'd contact Adobe or browse their forums.

[patrickjames]

Make a white Solid Color layer and set the blending mode to Color. Re-save the image and see if that fixes it.

[Brian K]

Patrick, that didn't seem to help, but here's something interesting. I imported the file into light room, and when I checked under white balance it said it had been saved "as shot", and then clicked on auto, in auto the color shifted on the screen image to the sepia coloration that it was printing as. Mind you, "as shot" looked fine, neutral, but under auto white balance it looked like the prints came out, seriously sepia.

[patrickjames]

If you are just looking to save the image Brian, maybe neutralize it in Lightroom then export it and see if that fixes it for printing. Perhaps the creation of a new file will get you around the corruption.