Scanner calibration & color casts

[Matus Kalisky]

I have asked about the differences between 35mm and 4x5 scanning targets Wolf directly and here is part of his answer:


Mr Wolf:
"... As there is no difference between
the dyes used in the film between various formats, you can use the
35mm targets with medium format and large format films. Actually,
all targets except A3 are produced on 8x10" films wich are than cut
to the wanted size. Roll films have a rather nasty behaviour of not
being flat when scanned in some scanners. The large format films are
thicker... no problems with bending in 35mm frames.
...
The large format film targets are less affected by scanner flare.
They are individually measurement. Combined this can result in a very
small but visible difference in scanned colors. "

I guess I will go with 4x5 - large enough not to get lost :-)

[Shutterbug]

I personally have tried both the 4x5 and the 35mm. In my experience I prefer the 4x5 because of the accuracy of the picture. 4x5 gives you a really crisp look at the image, which is all that matters to me.




_______________

25 yrs. professional wedding photography | Canon Mark IV | www.purplelionpaper.com

[mrladewig]

+1. Different films have different characteristics, and so require different calibrations.

+2

I don't care to calibrate my scanner+software to the film, but different films definitely have different interactions with the scanner. Kodak's films tend to scan much differently than Fuji even if they look very similar on the light table. Although most all the fujichromes scan similarly, they are each a bit unique as well. Scanning Astia isn't exactly like Velvia 50.

[neil poulsen]

Check out HutchColor.com's user guide.

http://www.hutchcolor.com/PDF/User_Guide.pdf

On page 15, they describe a method for adapting profiles for one type of transparency film, when the target is made from a different film. They also sell an excellent Velvia and Ektachrome targets that have over 500 patches.

Unfortunately, this method requires the use of a variable intensity, GTC Soft View style viewer. But, I thought it might be worth mentioning.

[Matus Kalisky]

Thank you all for replies and ideas - here is a short update:

I got the calibration slides (35mm) from Wolf and calibrated the scanner with all of them.

I went back to the images I have posted and re-scanned them. The cast got weaker, but the problematic reds were still there. I decided to take a less demanding slides for testing.

I chose an image taken on Provia 400X from Alhambra (see below) which was taken in a shade that was rather warm. The slide has nice nearly pastel color which is also rather neutral. The dynamic range of the image should really not be source of a problem. The scan produced image with slight greenish cast in lighter parts and with the nice brown color of the wooden door shifted towards strong red color - not nice.

Please note that the scan was of course made with an appropriate color profile of the scanner.
Please note that the images have the Adobe 1998 RGB color profile after scanning and were converted to sRGB before putting on the flickr.

1) So - here is a crop from the image that apart from simple levels adjustment (to lighten up the image as the scan was a little flat) had no additional adjustments:



2) Now after following additional adjustments I was able to come up with an image that up to some extent reflects the true colors in the slide. I managed to suppress the reddish cast in the wooden door, though it is a bit too yellow now (but still OK). The blue squares have a bit too much Magenta, but let's keep that aside for a moment:

The adjustments :
Selective color::Reds
-- Cyan +28%
-- Magenta -22%
-- Yellow +27%

Selective color::Yellows
-- Cyan +18%
-- Magenta: +14%

Color Balance::Midtones
-- Yellow/Blue +10
(Preserve Luminosity ON)

The result:



Status:
After some extensive color-adjustment work a rather reasonable scan can be obtained (still not perfect though). However I have to admit that I would expect better scans after calibration. Anybody has a good idea? I am slowly running out of mine ..

[Peter De Smidt]

What program did you use to make the profile? Did you use exactly the same settings to make the scan? When moving the file into photoshop, did you assign the scanner profile and then convert to your working space, noting that assign and convert have very specific meanings in Photoshop?

[Matus Kalisky]

I used the Silverfast Studio both for calibration and scanning. The Silverfast converts the scan to Adobe 1998 RGB before the file is stored so no conversion or assignment in Photoshop is necessary. On top of that the preview in Silverfast looks the same as the final scan when opened in Photoshop.

I have also tried (with Silverfast) changing the calibration profiles with the same preview - there were definitely changes and the correct one looked the best, but it still is far from what it should be.

I should probably mention that before the calibration I tried also the Microtek scanner Wizard and free version of Vuescan (can not use calibration in a free version). Silverfast and Microtek delivered both very similar casts and problems with red color. Surprisingly Vuescan was better.

I am considering getting a PRO version of Vuescan, but need to test the free version a bit more first (= understand the user interface). I need to figure out how the autofocus is handled (did 1 scan yesterday and it was rather unsharp compared to Silverfast)

[Ash]

None of my scans look 'right' regardless of profiling and calibration. I found my solution was scanning so that all the detail/tones were available (ie no blown out highlights) then using CameraRaw to post pro and correct colours and exposure if necessary.

[Matus Kalisky]

Ash - I agree that expecting perfect results out of the box is not realistic, however I do expect decently looking scans without blown-away gamut and strong color shifts. It is nearly impossible to get to "true" color in PP if the problems are too large.

[Tobias Key]

Have you tried CF systems colorperfect? You just do a linear raw scan, open it in photoshop, and run the plug in, there is trial version here

http://www.c-f-systems.com/Plug-ins.html

I've not used it with slides but it is a real life saver for colour negative and saves you loads of work. I has its own profiles for hundreds of films.