What technology is used to produce IT 8.7 Scanner Calibration Targets (e.g. those sold by Coloraid)? Are they produced photographically on slide film or digitally printed? Also, why separate targets are provided for different film stocks (Fuji Velvia/Provia and Kodak Ektachrome) if the goal is to calibrate the scanner, not the film?
How are IT 8.7 Scanner Calibration Targets made?
-
H
- Thread starter Romanko
- Start date
Recent Classifieds
-
Found Mamiya RZ67 120 film back
- Started by blee1996
-
For Sale Kodak 125PX Plus-X 35mm Film Lot (And More)
- Started by chuckroast
-
For Sale Large format Kodak 1443 Aerochrome color infrared film
- Started by MCB18
-
For Trade FT: My NEW 49mm IR 720 filter for your 52mm IR 720 filter.
- Started by AnselMortensen
-
For Sale FS: Mamiya RB-67 / RZ-67 Caps!
- Started by AnselMortensen
Forum statistics
I expect the film-based ones are used with some kind of old LVT/film recorder equipment. The paper-based ones can be made fairly simply by just outputting a digital file to any reasonable quality color printing service that outputs on chromogenic paper.
The gamut of each film is different and ultimately, the profile won't be just scanner-specific, it will be specific to the combination of scanner + film. So the notion that you calibrate the scanner is technically not entirely correct, in my view. You calibrate an imaging system, which consists of equipment plus media.
The gamut of each film is different and ultimately, the profile won't be just scanner-specific, it will be specific to the combination of scanner + film. So the notion that you calibrate the scanner is technically not entirely correct, in my view. You calibrate an imaging system, which consists of equipment plus media.
Yes, I am interested in transparent targets only.
So, if we shoot the same scene with two color reversal (slide) films and use the corresponding calibration sets the digital images should be identical (under ideal conditions of identical exposures and processing, of course). Is there a way to retain the film character and calibrate just the scanner part? If I am scanning some old Kodachromes I would definitely like the digital image to keep the unique look of this film. The same is true for Fuji Velvia.
The idea is that you calibrate the scanner + film as a combination, and that process is intended to retain fidelity of the color rendition of the original slide. So if you scan Kodachromes, you ideally calibrate your scanner with a Kodachrome target (if you can find one). The same for Velvia.
Velvia and Kodachrome will record a scene differently; this difference is preserved throughout the calibration. Keep in mind IT8 calibration is limited in scope to the medium plus the scanner; the original scene is not part of the calibration process. So:
no, that's not correct as far as I understand it. A calibrated Velvia scan and a calibrated Kodachrome scan from the same scene will come out looking differently in the digital domain just as the actual slides look differently, but the digital on-screen version should look virtually identical to the original slide in both cases. That's what you calibrate for.
The main question is how much utility there is in calibrating the scanner with targets on different types of film, since you'd theoretically expect that calibrating against a single type of slide should yield good results for other types of slide film as well. However, in practice, there will likely be minor differences that result from the unique ways in which different emulsions absorb and transmit light, for instance, which interacts in turn with the spectral sensitivity of the scanner.
Velvia and Kodachrome will record a scene differently; this difference is preserved throughout the calibration. Keep in mind IT8 calibration is limited in scope to the medium plus the scanner; the original scene is not part of the calibration process. So:
no, that's not correct as far as I understand it. A calibrated Velvia scan and a calibrated Kodachrome scan from the same scene will come out looking differently in the digital domain just as the actual slides look differently, but the digital on-screen version should look virtually identical to the original slide in both cases. That's what you calibrate for.
The main question is how much utility there is in calibrating the scanner with targets on different types of film, since you'd theoretically expect that calibrating against a single type of slide should yield good results for other types of slide film as well. However, in practice, there will likely be minor differences that result from the unique ways in which different emulsions absorb and transmit light, for instance, which interacts in turn with the spectral sensitivity of the scanner.
Why is the difference preserved throughout calibration?
By a "Velvia target," do you mean that an IT8 calibration target was photographed on Velvia? Because that sounds like it'd correct every different stock to look otherwise similar in color.
Or, as I think I'm gathering from your last paragraph, is it just an IT8 target printed on a slide that has been bleached of all silver, dye, etc; but still has all the layers that would exist in a developed slide? In other words, corrects the influence of everything that's not directly from the dyes, the "image itself", perhaps?
An IT8 target is typically not photographed. It's not like a Calibrite etc. color checker you photograph digitally for color matching purposes. The IT8 target on a slide typically won't even be photographed; it'll be recorded directly on film with an LVT etc.; see post #2.
The point of such a target is that the color densities on the target itself (in Velvia, Ektachrome etc.) are known and that you can adjust the output of the scanner in order to match those values. Thus, an IT8 target ensures that your digital output matches as close as possible the scanned media. It's not intended to create a close match to an original scene.
So to repeat: an IT8 target helps to get an output from the scanner that is as close as possible to the paper or film sample that is scanned. The idea of using one is that if you hold the original slide next to your (calibrated) computer monitor, the images on both will look more or less the same. This does NOT mean that the image on the slide matches the original scene photographed. Also, keep in mind that the IT8 target itself is not a real-world photographed scene. The IT8 calibration scope only comprises the slide (or paper) target and the scanner + its output profile. It does not extend to a real-world artefact such as a physical color checker that's being photographed.
No, it's a regularly processed slide.
I scan Velvia 50 and Ekrachrome and other chromes with Epson V600 and V850 scanners. Scanners don't change the way they scan. The hardware is fixed. All changes are made to the scanned file after the scan.
You can set these before the scan or scan flat with no scanner edits and make all changes afterwards to the scanned file using an editing program like Photoshop . Or you can use an inbetween approach as I do. The only setting I set for the scan is to set the black and white points just past each side of the data captures as shown in a prescan. I save the scan as a tiff. Then I make all further adjustments in Lightroom.
Or you can scan flat and adjust the black and white points (levels) in Lightroom. 90% of the colors are recaptured to match the slide with just this one setting. Then tweak in your editing program to fine tune. The colors will match the original colors for the most part. I never compare the resultant image to the slide. My thinking on this is it's my photograph and my creation. The final color do not have to match the colors of the slides designed by some dead Japanesee 50 years ago. As long as they are acceptable to me, that's fine. The end result will pretty much match the slide in any case. I never used IT targets so I dont know how that works.
It’s worth being in mind that Kodachrome IT8 targets are very rare and cost €800 or so, so this may be an academic excercise.
This makes sense since E6 and Kodachrome are different in every respect: process, dyes, contrast, characteristic curves, you name it.
Coloraid IT8.7 targets are available for a variety of E6 film stocks (Ektachrome, Velvia, Provia etc.). Each target is calibrated at a batch level (if not individually). I am struggling to understand why multiple film stocks. Shouldn't one be sufficient to calibrate the scanner provided the XYZ / Lab values of the target patches are known? Does it even need to be a photographic process?
Yes, my interest is mostly academic. I have no need to produce colorimetrically correct scans of Kodachrome slides.
And so is mine. However, sometimes I scan historical slides made by other photographers and I would like my output to be at least consistent, if not accurate.
I see where you're getting stuck, and yes, at face value, you'd expect it to work the way you say. Calibrate the scanner with whatever transparency material you've got that has some color patches with known densities on it, and you should be good. At least - in the ideal world. But of course, the real world works differently. In reality, a magenta patch on two different pieces of slide film that appear more or less the same color to the human eye may in fact have subtly different transmission densities. Moreover, the scanner's sensor sites (typically CCD) have their own spectral sensitivity curves. Both of these things interact with each other, and this can result in (usually) small deviations. This is what you calibrate for.
For instance, here are the spectral dye density curves for Ektachrome 100 (magenta curve) and Velvia 50 (cyan curve) overlayed:
Notice how they follow a similar pattern (indeed, it looks like Kodak and Fuji really are using pretty much the same dyes), but there are subtle differences. Note for instance the shift in the magenta peak and a subtly different gradient in the left-side of the cyan curve.
If you were to put the sensitivity curve of the scanner's sensor on top of such a plot, you will also see that certain wavelengths will be emphasized while other are suppressed, and that there's a truckload of channel crosstalk going on. Part of this is filtered out by the scanner's software/firmware on the basis of calibration procedures that are done during product development (which would boil down to pretty much what you were saying). But part of it will ring through in the final scans because there's no way to perfectly control for the variety of combinations of transparency film and sensor behavior using a single calibration setting. Hence, to squeeze the last bit of accuracy from the system, you can profile the scanner against an IT8 target that's made on the film you're going to scan lots of.
If you don't use this IT8 profiling, you're thrown back to the factory/product-inherent calibration of transparency targets, which for most purposes will in practice be good enough.
I believe we can exclude the color perception of human eye from the equation. It comes into play at a later stage when viewing the digital image on a computer screen, or a print on transparent media on a light table. Both presentations introduce a whole lot of their own transformations to the image so let's focus just on the transparency to digital image transform.
Till this point, it's all good. To rephrase: (1) we have spectral sensitivity curves for the three color channels of the scanner which most likely differ from that of the dyes in the transparency; (2) the scanner applies some built-in corrections to compensate for this discrepancy (color space transformation?); (3) this compensation is not ideal and can be improved by using a IT8.3 calibration target. It has enough color patches to build a fairly detailed 3D Look-up Table (LUT) that can be applied to a scanned transparency.
Calibration targets for different film stocks and batches of the same film have different XYZ/Lab values of the color patches. My understanding is that this difference would be calibrated out and the resultant 3D LUT will be the same independent of the film stock. Obviously, due to calibration uncertainty there will be slight differences in the LUTs built using, say, Ektachrome and Velvia.
If I understood you correctly, you are saying that if the same film stock is used to make the calibration target and the transparency, the calibration uncertainties will be smaller and the scanned digital image will be more accurate. Intuitively, I believe you are right, but I can't explain why this would be the case.
What @koraks is describing is called
Unfortunately, we can not.
Colour materials (slide film in this case) are made for humans. CCD sensors work differently than human colour vision. So a human can see two distinctly different hues on two different materials and the scanner sees them as one hue (it can't distinguish between the two). Or vice versa. This is called scanner metamerism problem.
Profiilng for different materials reduces the problem.
Unfortunately, we can not.
Colour materials (slide film in this case) are made for humans. CCD sensors work differently than human colour vision. So a human can see two distinctly different hues on two different materials and the scanner sees them as one hue (it can't distinguish between the two). Or vice versa. This is called scanner metamerism problem.
Profiilng for different materials reduces the problem.
Using levels gets 90% of the original colors. Here's a comparison of Ektachrome and Provia I did using 4x5 film. You can see how Ektachrome greens are greener and its reds are redder than Provia's orangey look.
Attachments
Yes; I only mentioned it to illustrate that differences in color rendition can go beyond what we can easily observe - either with our own eyes, or by technical means.
Yes, that sums it up fairly well.
Indeed, it's essentially a metamerism problem.
| Photrio.com contains affiliate links to products. We may receive a commission for purchases made through these links. To read our full affiliate disclosure statement please click Here. |
PHOTRIO PARTNERS EQUALLY FUNDING OUR COMMUNITY:  |