Page 1 of 4 123 ... LastLast
Results 1 to 10 of 31

Thread: Scanner calibration & color casts

  1. #1

    Join Date
    Sep 2005
    Location
    Germany, Aalen
    Posts
    849

    Scanner calibration & color casts

    After coming back from vacation together with quite some film to be scanned I am back to an old problem of mine with my scanner Microtek F1.

    The problem:
    slides have strong cast towards magenta in middle tones and even more in shades. Strong post processing is needed to remove it.

    Short tech part:
    I run Microtech F1 with Silverfast Ai with MacBook (10.5.8). I use IT8 target (Kodak Ektachrome) that came along with the Silverfast. The calibration procedure is fast.

    For the scans below - all were done on 6x6 Velvia 100 slide with NO in scanner adjustments (apart from the calibration, obviously).

    All images are clickable to get larger versions.

    I use a Velvia 100 6x6 slide to show the problem, but I have experienced this also with Provia and Astia.

    If I scan the calibration slide itself AFTER the calibration - everything looks fine. Please note that apart from resizing and converting to sRGB there was no post processing applied. The conversion from Adobe1998 to sRGB introduces differences that are negligible to the problem discussed.



    ________
    Now here is a sample scan without any adjustments:



    ________
    The same scan after making it a bit lighter (global level adjustment) to make the problem more obvious:



    ________
    After basic color-wise level adjustment trying to get closer to that how the slide actually looks:



    But there is one more problem that comes along - namely the red colors tend to get step-wise extremely red - like the red channel would run out of color space or something. The struck areas are suddenly much more red than should be and look very unnatural. This problem can also be seen in deep shades

    This is not the best slide to show this problem, but just trust me - it is there without ANY post processing and can look VERY nasty if there are some truly red objects (like a red sofa or human skin shot under bulb light)


    CONTINUED on first answer ....
    Matus

  2. #2

    Join Date
    Sep 2005
    Location
    Germany, Aalen
    Posts
    849

    Re: Scanner calibration & color casts

    ....

    ________
    Now a few 100% (2400 spi) cut outs to show the color posterization problem:



    _________
    My observations:

    I get these problems with both SilverFast and Microtek Scan Wizard. I got somewhat better results with Vuescan. The issue of color cast and color posterization seem to be related.

    __________
    I would appreciate ANY help or ideas how to solve this problem as both the color cast and the color posterization make for scans that are barely usable even after lot of PS work - especially if there are shadows in the photo.
    Matus

  3. #3

    Join Date
    Mar 2005
    Location
    Aurora, Colorado
    Posts
    526

    Re: Scanner calibration & color casts

    I'm using an Epson 4990 and Epson Scan software. I found Velvia difficult to scan before I made a profile for it. I scan without any color management and when I open an image in PS I apply the profile and convert to my working color space (Adobe RGB). This usually works well but Velvia can easily blow out the red channel. Sometimes I have to desaturate the reds by quite a bit.

    Some time ago I learned an easy technique for cleaning up color casts. It works like this:
    1. Duplicate the background layer.
    2. Run Filter->Blur->Average on the duplicate layer.
    3. Add a Levels or Curves adjustment layer. In the adjustment dialog box, select the gray point eyedropper and click anywhere in the image.
    4. Trash the duplicate layer.

    This technique isn't always perfect but the example image you posted responds well to it. I applied it to your second version - the one with the strong magenta cast.

    Hope that helps.
    Never is always wrong; always is never right.

    www.LostManPhoto.com
    www.MarkStahlkePhotography.com

  4. #4
    Peter De Smidt's Avatar
    Join Date
    Jan 2001
    Location
    Fond du Lac, WI, USA
    Posts
    8,944

    Re: Scanner calibration & color casts

    The problem might be that you are using an ektachrome target to make a correction file for velvia. I suggest that you buy Wolf Faust's inexpenive velvia IT8 target and use that calibration when scanning Velvia.

    I have Wolf Faust's whole set of IT8 targets. They give very different results.
    “You often feel tired, not because you've done too much, but because you've done too little of what sparks a light in you.”
    ― Alexander Den Heijer, Nothing You Don't Already Know

  5. #5
    Resident Heretic Bruce Watson's Avatar
    Join Date
    Sep 2003
    Location
    USA, North Carolina
    Posts
    3,362

    Re: Scanner calibration & color casts

    Quote Originally Posted by Peter J. De Smidt View Post
    The problem might be that you are using an ektachrome target to make a correction file for velvia.
    +1. Different films have different characteristics, and so require different calibrations.

    Bruce Watson

  6. #6

    Join Date
    Sep 2005
    Location
    Germany, Aalen
    Posts
    849

    Re: Scanner calibration & color casts

    Quote Originally Posted by Peter J. De Smidt View Post
    The problem might be that you are using an ektachrome target to make a correction file for velvia. I suggest that you buy Wolf Faust's inexpenive velvia IT8 target and use that calibration when scanning Velvia.

    I have Wolf Faust's whole set of IT8 targets. They give very different results.
    This is something I was considering could be the reason. Still - at least in theory - if the the scanning slide covers sufficient color range be independent from what color slide one used? If not that it should be only the color of the "base" that cases an offset. Or am I completely wrong?

    But anyhow I will give a try to new set of calibration slides.
    Matus

  7. #7
    Moderator
    Join Date
    Apr 2009
    Location
    Northern Virginia
    Posts
    5,614

    Re: Scanner calibration & color casts

    Quote Originally Posted by Matus Kalisky View Post
    This is something I was considering could be the reason. Still - at least in theory - if the the scanning slide covers sufficient color range be independent from what color slide one used? If not that it should be only the color of the "base" that cases an offset. Or am I completely wrong?
    You would think that for transparencies it wouldn't matter that much--they are all designed to look "realistic". But it does. Ektachrome has always had a reputation for a blue/green cast, and the calibration is correcting for that cast by applying yellow and magenta (read: red). Unfortunately, those colors are strong in Velvia already. It's like looking through a filter laying on a white card--you get double the color and that makes the shift quite obvious.

    Rick "who has also bought a few calibration targets from Faust" Denney

  8. #8

    Join Date
    Sep 2005
    Location
    Germany, Aalen
    Posts
    849

    Re: Scanner calibration & color casts

    I have one question left. Will 35 mm slide be large enough for good color calibration or should I go with 4x5" targets? I mean - the spacial resolution of my scanner is not as good as dedicated 35mm scanners and therefore the edges between different colors will get "smeared" during the calibration procedure. And as I actually need only 1 calibration target (lucky me - it covers Astia, Provia (100 & 400x) and Velvia 100) the additional expense for 4x5 over 35mm is not that much more. So I am wondering ...
    Matus

  9. #9
    Peter De Smidt's Avatar
    Join Date
    Jan 2001
    Location
    Fond du Lac, WI, USA
    Posts
    8,944

    Re: Scanner calibration & color casts

    I use the 35mm targets with no problems on a flatbed, but I believe that Mr. Wolf's 4x5 targets are individually measured. If so, that'd be an advantage for the larger piece of film.
    “You often feel tired, not because you've done too much, but because you've done too little of what sparks a light in you.”
    ― Alexander Den Heijer, Nothing You Don't Already Know

  10. #10
    Resident Heretic Bruce Watson's Avatar
    Join Date
    Sep 2003
    Location
    USA, North Carolina
    Posts
    3,362

    Re: Scanner calibration & color casts

    Quote Originally Posted by Matus Kalisky View Post
    I have one question left. Will 35 mm slide be large enough for good color calibration or should I go with 4x5" targets?
    Large enough? Yes. But you should know that while the emulsion is largely the same, the film base is not. There are differences between the 35mm version of a film and the 5x4 version of the same film.

    Is the 35mm target "good enough" to profile 5x4 film? That depends on how accurate you want your profile.

    The real problem as I see it (and I realize that not a lot of people see it the same way I do), is that not that much film that comes to a scanner hits the profile on the nose. That is, has the same black and white levels, and the same contrast index, was exposed with the correct color temperature light, etc. So you'll always have *some* level of corrections to make in post.

    What a profile does is make it somewhat easier for a novice scanner operator (that is, the scanner settings have to replicate those used when creating the profile [or you invalidate the profile] so there are fewer choices to make). It only works because tranny film will nearly always come out relatively close to the target -- it has to or it doesn't function well as a display transparency. When projected, it has to give a convincing black and a reasonable white and have reasonable contrast.

    But unless the tranny is exposed in the studio with the correct color temperature lighting and controlled dynamic range, it's not going to be a *very* close match to the film used to make the profile. And thus there will likely have to be some color correction done in post.

    So... it devolves down to *how much* color correction work are you willing to do.

    A good scanner operator who really understands the software can often get as close as a profile for "average" films. And for films that were exposed under difficult light and conditions (think landscape photography), can often do better. So *not* using a profile can sometimes give *better* results and *less* work in post.

    Some people swear by scanner profiling. Some people swear by learning their software and their scanner. Just something to think about.

    Bruce Watson

Similar Threads

  1. Scanner Calibration
    By AnselAdamsX in forum Digital Hardware
    Replies: 0
    Last Post: 31-Oct-2006, 18:48

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •