Page 1 of 1

IT8 calibration target neutrality?

PostPosted: Tue Sep 15, 2015 7:48 am
by Rich P
I've just calibrated a new Plustek 8200i with the included Fuji target on my Mac Pro (which has an X-Rite calibrated screen) and when I scan the target itself it appears to have a cool/bluish cast. When I open the resulting image in editing software the RGB values of the 'greyscale' section at the bottom of the slide sure enough show a lower value in the red channel.

I know it is only a target, and I haven't had chance to scan a 'real-life' slide yet, but the colours look a bit unpleasant which is not what I see when I hold the slide to the light.

My question is: should the slide give neutral/white-balanced values for the black to white areas or is a bluish cast expected on this slide when calibrated correctly? Thanks you!

Re: IT8 calibration target neutrality?

PostPosted: Fri Sep 18, 2015 1:33 pm
by LSI_Ketelhohn
Dear customer,

The greyscale on the target is only a reference it is not truly neutral and does not need to be.

Kind regards,
Arne Ketelhohn.

Re: IT8 calibration target neutrality?

PostPosted: Sat Sep 19, 2015 4:27 pm
by Jossie
Hi Rich,

may I try to answer your question in somewhat more detail?
Should the slide give neutral/white-balanced values for the black to white areas or is a bluish cast expected on this slide when calibrated correctly?

To know the proper appearance of the grey patches one has to look into the reference file coming with the target. Relevant to your question are the a- and b-values for patches labelled GS0 (white) to GS23 (black). These a- and b-values specify the colours of the light source used to expose the target slide. What you aim at with the whole calibration process is, that on your calibrated monitor the patches appear with these colours as close as possible. The original colours of the light source are modified by the colour response of the Fuji film and the colour response of the scanner. Both of these have to be calibrated out via the ICC profile created from the IT8-target to achieve this goal.

For most targets I know the GS patches are of neutral grey, i.e. their a- and b-values are close to 0 (this is not mandatory but helpful. Ignore the L-values, they specify the luminance.). If the human eye should not be able to discern any colour cast in these grey patches the quantity dE calculated as dE = srqt(a^2+b^2) should be less than ~1.

An easy way to calculate the dE values is to load the reference data into EXCEL and do the calculation there. Here are the data for one of my Fuji targets:

Code: Select all

SAMPLE_ID       L        a        b    dE
GS0         90.99     0.16     1.60  1.61
GS1         80.88     0.02     0.67  0.67
GS2         77.18     0.03     1.31  1.31
GS3         73.18     0.23     0.73  0.77
GS4         69.27    -0.63     0.55  0.84
GS5         65.26    -0.59    -0.05  0.59
GS6         61.06    -0.04     0.38  0.38
GS7         57.18    -0.42     0.49  0.65
GS8         52.97    -0.07     0.00  0.07
GS9         48.91     0.04    -0.19  0.19
GS10        45.06    -0.55    -0.10  0.56
GS11        41.01    -0.60    -0.01  0.60
GS12        36.98    -0.39     0.09  0.40
GS13        33.02    -0.51    -0.35  0.62
GS14        28.99    -0.26    -0.12  0.29
GS15        24.78    -0.28    -0.15  0.32
GS16        20.96    -0.34    -0.03  0.34
GS17        17.27    -0.66     0.44  0.79
GS18        13.20    -0.53     0.95  1.09
GS19         9.36    -0.43    -0.30  0.52
GS20         5.53    -0.46     0.31  0.55
GS21         3.49    -0.33    -0.17  0.37
GS22         1.84     0.05    -0.07  0.09
GS23         0.85     0.18    -0.14  0.23

Therefore the grey patches for this target should appear on my calibrated monitor as really neutral grey -- and they do. The target slide, of course, if viewed on a light table, will look different and the greys have a colour cast introduced by the colour sensitivity of the film. For one of my other Fuji targets the a-values in the reference file are in the range -1 to -3, b-values scatter around 0. Thus I would expect a very, very slight greenish cast in the grey values.

You should measure the a- and b-values for your calibrated scan of the target. As there seems to be an issue with the the Lab-values displayed by the densitometer in SilverFast8, I would use SilverFast6 HDR or RawTherapee to measure the patches and check, if their dE value is indeed as specified by your reference table, on average.

Hermann-Josef