Kodak grey card scan - green colour cast?

[brian_mk]

Epson Perfection 4870 Photo.
Silverfast Ai Studio 8.8r25
Windows 10 64bit.

I calibrated the scanner using an IT8 Target from Wolf Faust.
The Silverfast IT8 software created a new Input icc profile:-

When I scan the IT8 target, the colour balance looks fine when viewed on my Eizo calibrated monitor.
There is no color cast evident in the neutral grey border.

If I try a scanning a Kodak grey card at the same time, the grey card has a very obvious green cast.
When the originals are viewed next to each other in daylight, the greys both appear virtually idenitical.
The scanner apparently registers the light reflected from the surface of the grey card differently to the human eye.

If I measure the grey card using a refelective densitometer, I get almost equal readings for the R,G,B densities.

I have also tried using the Epson scanning software (using it's own Epson input profiles). I get a similar result.

Can anyone explain why the scanned grey card image looks so different to the original?
I can understand why there could be a slight difference - matt vs glossy surface etc. but I did not expect it to look this bad.

 

[ced]

One possibility: https://www.sciencedirect.com/topics/engineering/metamerism

 

[brian_mk]

Yes it could be down to metamerism.

https://en.wikipedia.org/wiki/Metamerism_(color)

If it is, then it seems to me that trying to create in input icc profile for the scanner is a waste of time.
You would need different profiles for every type of printed source material that you scan.

So in my case the scanner may be calibrated for scans made using Fujicolor Crystal Archive paper (the same as the IT8 target) but anything else could give completely different results.

I see a similar effect if I make a print from the scanned IT8 target on Epson Premium Glossy paper and re-scan the printed version.
To my eyes, when viewed in daylight, there is no obvious colour cast on the printed version when compared to the original.
However, when scanned, the printed version has an obvious green cast.

I guess this is why you should not try to use a flatbed scanner to create or check icc profiles for a printer.
Presumably a spectrophotometer can make measurements that more closely match what the human eye sees.

 

[MattKing]

There is a reason that Kodak grey cards were intended to be set at an angle (20 degrees, IIRC), rather than parallel with the subject plane.
Does your scanner use a fluorescent source?

 

[wiltw]

Brian, I just did two scans and evaluated the neutrality via the eyedropper tool within Lightroom, as a basis of comparison with you result:

1. Scan of MacBeth Color Checker card
2. Scan of genuine Kodak 18% grey card

• LR evaluated #1 gray patch #4 Before adjustement as R48.3, G48.5, B47.4 initially,
  and After using eyedropper tool to render neutral the readings were R48.8, G48.7, B48.5
• LR evaluated #2 card Before adjustement as R47.6, G48.5, B49.2 initially,
  and After using eyedropper tool to render neutral the readings were R49.3, G49.4, B49.3, no

These values were obtained with Canon 8800F flatbed scanner, using NO calibration adjustment, nor any IPS monitor adjustment.
These results prove that Kodak grey cards do not ordinarily deviate much from other standards like the ColorChecker card.

My suggestion, since your process has both scanner calibration and monitor calibration, is to scan the gray card to JPG and interrogate the values of R-G-B with a postprocessing program's Eyedroipper tool and look at the values...they should all be very similar in range, all be very close to 50% (mid-tone).
The neutrality of interrogated values will bear witness to proper calibration of the scanner, and viewing that image on your monitor should similarly show absolute neutrality of presentations, and printing that same image should result in neutrality of the printed image.

 

[brian_mk]

My suggestion, since your process has both scanner calibration and monitor calibration, is to scan the gray card to JPG and interrogate the values of R-G-B with a postprocessing program's Eyedroipper tool and look at the values...they should all be very similar in range, all be very close to 50% (mid-tone).

I had already tried that.
The measurements in Photoshop confirm that the green value is high for both the grey card and the printed copy of the IT8 target.
The scanned original IT8 target is pretty much perfect (using the Silverfast calibrated icc profile for the scanner).

BTW The Epson 4870 uses a cold cathode flourescent tube.
Could it be that the Canon 8800F uses a different kind of light source for reflective?

I would not expect Kodak grey cards to be super accurate for white balance - they were mainly intended to get correct exposure.
I would expect the Macbeth color checker to be considerably better - especially given the high price.
What does surprise me is the difference in colour cast between the original and the scanned versions of the Kodak card.
As already mentioned, the Kodak card appears completely neutral when measured with a Macbeth densitometer and when viewed in daylight.

 

[wiltw]

Brian,
Since your scan and the print both have the greenish tint, I would suspect the validity of your scanner calibration efforts! Can you remove that calibration file, to restore your system to 'stock' and try with an uncalibrated system to begin with and verify results with Eyedropper tool and print, then redo the calibration and check that result with Eyedropper again?

I have witnessed too many folks screwing up things in their calibration methodology, to know that procedural errors do occur!

 

[brian_mk]

There is a reason that Kodak grey cards were intended to be set at an angle (20 degrees, IIRC), rather than parallel with the subject plane.

That's more likely to prevent specular reflections from the surface giving false exposure readings.
I have noticed this can be a problem when measuring light refelected from a grey card in sunlight using a lightmeter.

 

[brian_mk]

Brian,
Since your scan and the print both have the greenish tint, I would suspect the validity of your scanner calibration efforts! Can you remove that calibration file, to restore your system to 'stock' and try with an uncalibrated system to begin with and verify results with Eyedropper tool and print, then redo the calibration and check that result with Eyedropper again?

I have witnessed too many folks screwing up things in their calibration methodology, to know that procedural errors do occur!

Yup - already thought of that.
I tried using the Epson scanner software instead of Silverfast. That uses Epson's own profiles for the scanner.
I get pretty much the same result.
The IT8 target scan looks ok. The grey card scan looks green.

 

[brian_mk]

BTW if I adjust the colour balance in Photoshop to get the scanned grey card to appear neutral as measured by the eyedropper, then the scanned IT8 target scanned at the same time appears magenta as you would expect.

All this makes me convinced it has to be a metamerism effect and nothing to do with scanner icc profiles.

 

[wiltw]

Yup - already thought of that.
I tried using the Epson scanner software instead of Silverfast. That uses Epson's own profiles for the scanner.
I get pretty much the same result.
The IT8 target scan looks ok. The grey card scan looks green.

Then I don't know what to say, apart from the fact that I can scan a Kodak card and get as neutral results as scanning a MacBeth card, using a 'stock' system. Maybe the light source of the scanner is to blame.
Checking the web about Canon 8800F light source reveals:
Scanner, Model Element, Light Source
8800F, CCD, White LED - Platen CCF lamp - Film

 

[brian_mk]

Then I don't know what to say, apart from the fact that I can scan a Kodak card and get as neutral results as scanning a MacBeth card, using a 'stock' system. Maybe the light source of the scanner is to blame.
Checking the web about Canon 8800F light source reveals:
Scanner, Model Element, Light Source
8800F, CCD, White LED - Platen CCF lamp - Film

That's interesting - your Canon uses LED for reflective. My Epson uses CCF. Could that be why the results we get differ?

 

[BrianShaw]

That's interesting - your Canon uses LED for reflective. My Epson uses CCF. Could that be why the results we get differ?

Does what Kodak says about color balance in motion picture usage of gray card give any clues?

https://www.kodak.com/en/motion/page/gray-card

 

[brian_mk]

The human eye combined with the brain plays tricks.
If we look at a light source containing pure red, green and blue light, we see what we perceive as white.
A spectrophotometer will see and measure the separate red, green and blue wavelengths.
This is why we get metamerism effects.
A scanner / photodiode / ccd / sensor sees light differently to the human eye.
Evolution has made the human eye more sensitive to green since most of the light we see lies within that range.
It's a facinating subject.

 

[brian_mk]

It's also easy to get 'bamboozled' by companies that try to sell you expensive solutions to CMS problems that don't really work.

 

[nmp]

I got identical results (green cast) from the IT8 Target from Wolf Faust. I stopped using it and went back to native calibration.

 

[brian_mk]

Some more results from this morning...

First a scan of some colour charts using the Silverfast icc profile from Wolf Faust.
They all have a green cast.

 

[brian_mk]

Same setup but but with a Kodak gray card.
The green cast is still apparent.

 

[brian_mk]

Now I reverted back to the native icc profile that came with the scanner (named EPSON Perfection4870).
The green cast has gone!

 

[brian_mk]

It looks like Niranjan is right. The problem must be with the reference file supplied with the Wolf Faust IT8 target.

I'm not sure why I get the green cast when using the Epson Scan software rather than Silverfast.
The Epson Scan software does not show which input profile is actually being used.
I tried it again this morning. The Epson Scan still shows a green cast. (See attachment).

 

[nmp]

It looks like Niranjan is right. The problem must be with the reference file supplied with the Wolf Faust IT8 target.

I'm not sure why I get the green cast when using the Epson Scan software rather than Silverfast.
The Epson Scan software does not show which input profile is actually being used.
I tried it agian this morning. The Epson Scan still shows a green cast. (See attachment).

Nice to know your results seem to corroborate mine. I thought there was something wrong with my procedure and left it that. I strictly use Vuescan and my scanner is Epson 3200. With the default profile I was getting the images that looked a little warmer to me so I thought may be I need calibration. After the green cast issue I decided I can just fix the warmth issue in Photoshop using a correction curve and leave IT8 target to the side.

I wonder what Wolf Faust might have to say about this.

:Niranjan.

 

[PhilBurton]

[QUOTE="
I wonder what Wolf Faust might have to say about this.

:Niranjan.[/QUOTE]
Ask him. wfaust@coloraid.de

 

[brian_mk]

I'm still investigating these issues - although things are delayed because I'm having problems with the fan on my graphics card.
I can't find a replacement fan in the UK so I've had to order one from China. The GPU still runs but I'm afraid to use it much until the fan issue is sorted.
In the mean time I'm using an old Win 10 laptop.
The laptop has a pretty awful screen so the images I included in earlier posts don't show the up the green cast in the same way as the Eizo monitor on my desktop PC.

I have made some new discoveries however...

First I found that the Wolf Faust IT8 target comes with two different reference files: One measured on a white backing, the other on black.
As the 4870 uses a white backing in reflective mode, I went back and re-ran the Silverfast IT8 calibration process using the white reference file.
It made little or no difference to the green cast.

Next I installed Epson Scan on the laptop and connected the Epson 4870 scanner.
This was a clean install with the latest software from the Epson website.
Only one input profile is available named 'EPSON Standard'. This is also the case on my desktop PC.
The name 'EPSON Standard' doesn't correspond to any of the icc profile files in C:\Windows\System32\spool\drivers\color.
I was initially unable to locate the profile I was using with Silverfast on the desktop PC named 'EPSON Perfection4870'.
This is the profile that appeared to get rid of the green cast.
I subsequently discovered this profile is actually installed as part of Epson Scan. There are two icc files named prer487_r.icm for reflection and and per487_t.icm for transmission. I assume these are the profiles used when you select 'EPSON Standard'.

Other icc profiles on the desktop PC that relate to Silverfast and the 4870 all have an 'SF' prefix in the filenames.

Since Epson Scan only appears to work with it's own internal profile, that could explain why it still gives a slight green cast (although not as bad as Silverfast when using the profile from the WF IT8 Target).

Another discovery is that Epson Scan does not recognise icc v4 profiles. It only works with v2. This could be why Epson Scan on the desktop PC does not allow me to select any of the Silverfast 4870 icc profiles: They do not appear in the pull down list.

 

[brian_mk]

More wierd results...

Today I ran Silverfast on my desktop PC.
I used the WF IT8 profile and carried out a prescan of the IT8 target and Kodak grey card as before.
I used the pipette tool, Selected 'White Point' in the pull down menu and clicked on the greyscale white rectangle on the IT8 target to set the white point.
Then I did the same on the black rectangle to set the black point.
The obvious green cast on the grey card then appeared.

So then I selected 'Reset Automatic' in the pipette tool pull down menu to undo the white & black points.

Instead of using the pipette tool, I then adjusted the slider controls in the histogram tool to set the shadow and highlights manually.
Now the green cast does not appear!
It seems that setting the white and black points using the pipette tool affects the colour balance as well as the luminocity and produces the green cast.

I repeated the above but this time also set the 'Neutral Point' in the pipette tool pulldown.
I clicked on the grey card image for the neutral point.
The effect was to neutralise the green cast.

I'm not sure what happens as far as the icc profile is concerned when you do this.
It looks to me as if the icc profile that was active when the prescan was carried out gets overridden.

All I am trying to do is is to get a reasonably neutral looking scan without having to manually tinker with the colour balance every time.
Why is that so hard? I am close to giving up!

 

[Mr Bill]

Why is that so hard? I am close to giving up!

Well, it's a very complicated thing going, as you mentioned about the metameric issues. And I truly don't know why the software is not more seamless. I would hazard a guess that there is just not enough money to be made, so development is done by just a small number of people.

I've been something of a color guy, not so much anymore. And not much experience with scanners. But I can perhaps make some observations.

You mentioned not being able to find (or use) certain ICC profiles. In case you don't know, ICC profiles also have an "internal name" in addition to the filename. And in my experience software generally uses the internal name. So it's possible, when you can't find a particular profile, that you are looking for the wrong name.

Regarding your Wolf Faust target, I note that it says it's on Fuji Crystal Archive, which is (presumably) an RA4 process paper. These papers don't have a flat spectral response, by which I mean that, if you took readings with a spectrophotometer, you would not get a flat line on a graph. It would have some wiggles in it. So you might figure that a scanner, looking at narrow spectral ranges, is gonna see different "colors," depending on exactly where it looks. So this is why you might need the special test target, to instruct your specific scanner exactly how it should "see" a certain set of readings. Once so "tuned" it should be able to correctly interpret that particular type of print media, BUT NOT NECESSARILY OTHERS.

Let me show, as an example, an image I had posted elsewhere. It shows an example of a color negative FILM (not paper), in a "neutral" area, and roughly how a specific densitometer response (Status M, used for "process control monitoring" of such films) "sees" that film. So you can clearly see that if the densitometer were to look at a slightly different spectral zone, it would produce different readings. Let me emphasize that this is more extreme than a photo PAPER, but the situation is fundamentally the same.

Lastly, with respect to Kodak gray cards, I think that they often get an undeserved bad rap. I've personally looked at more than a handful, using a spectrophotometer, and all I've looked at had a pretty decent spectral response. And what I've measured correlated pretty well with the graphs included in the package. Certainly not enough off to give the casual user problems. Also, something that doesn't seem very well-known, at least on the internet, is that, for some years, Kodak's gray cards have been using the same Munsell technology as the Macbeth Color Checkers. I don't know how long this has been the case, but I'm guessing roughly on the order of ten to twelve years; maybe someone has better info?

Hopefully this will shed a little light on the situation.