+ Reply to Thread
Results 1 to 6 of 6

Thread: Flatbed Scanner instead of Print Densitometer

  1. #1
    Donor Gargoyle will become famous soon enough
    Join Date
    Dec 2010
    Posts
    70
    Thanks
    5
    Thanked 72 Times in 29 Posts

    Flatbed Scanner instead of Print Densitometer

    If I profile a flatbed scanner to Q60 color targets I can literally turn this into a Lightness densitometer and measure L values more quickly while still fullfilling ISO 12647?

  2. #2
    Junior Member Ray Vega is on a distinguished road
    Join Date
    Oct 2012
    Posts
    10
    Thanks
    0
    Thanked 2 Times in 2 Posts

    Yes and No

    Densitometers have filters in them, often many different filters, to match the measurement instrument to different types of media (status A, status M), since different media behave differently (spectral output curves define this behavior). If unmatched, you wind up with the a reading telling you say, that you have (x) density for Magenta, but the instrument being unmatched is not seeing just Magenta, it's also seeing a little bit of Cyan. This is a simplified explanation, but filtration is the first of several issues you'd face if you're after accuracy (other issues include insensitivity in low signal areas leading to noisy readings, drift and inconsistency). Having warned you, there was an article in Creative Darkroom magazine years ago on linearizing and calibrating flatbed scanners to use them as densitometers. Used in a closed loop situation with unchallenging signal ranges (flatish originals in tone and gamut), you may find yourself achieving what you need to your satisfaction. But, I wouldn't in this case call the scanner a densitometer as your density numbers are likely not going match actual densitometers very well. Finally, what's the filtration of the scanner?

  3. #3
    Donor Gargoyle will become famous soon enough
    Join Date
    Dec 2010
    Posts
    70
    Thanks
    5
    Thanked 72 Times in 29 Posts
    Thanks for your advice on this. I agreed very well. That's why I am attempting to use ICC method to characterize the scanner with unknown filtration and mapping the RGB values to densitometric values of a reference target. I also understand about metamerism factors and that's why its application will be limited for use only to which it was profiled. Regarding scanner noise and dynamic range issues IMO the scanner can see far more dark density details than printed paper media can attain in its darkest densities? Besides we can also use multipass and use 16 bits scanning. I'm only wondering about the accuracy it can achieve by using this method?

  4. #4
    Donor Gargoyle will become famous soon enough
    Join Date
    Dec 2010
    Posts
    70
    Thanks
    5
    Thanked 72 Times in 29 Posts
    My logic is this. If scanners can be ICC characterized for film and reflective targets why not status densities as well?

  5. #5
    PRC Member spider is a name known to all spider is a name known to all spider is a name known to all spider is a name known to all spider is a name known to all spider is a name known to all
    Join Date
    Nov 2010
    Posts
    1,342
    Thanks
    126
    Thanked 626 Times in 356 Posts
    Quote Originally Posted by Gargoyle View Post
    My logic is this. If scanners can be ICC characterized for film and reflective targets why not status densities as well?
    because your scanner doesnt have a standard d50/d65 light source.

    the spectrum of your scanner light is not defined and the scanner light has not exactly the same intensity over the whole scanning area.

    also the degree of the light source and the ccd is not the same as for spectrophotometers.

    non existing uv-filter might be another problem.

  6. #6
    Donor Gargoyle will become famous soon enough
    Join Date
    Dec 2010
    Posts
    70
    Thanks
    5
    Thanked 72 Times in 29 Posts
    Okay and acknowledge. Theoretically however this still looks feasible. A typical densitometer lets say has DMax of 2.5D with +-0.01D accuracy. That to me is only 8 bits fidelity. Now for flatbed scanner the principle is still RGB albeit different filters and light sources. That will have lower sensitivity of course but will not be also close to zero because of the former. Now lets assume the sensitivity is merely 50% compare to a real densitomter or do you think the sensitivity can go much lower than that?

    Now the bottom line, since the flatbed sensitivity is only 50% (assuming) then mathematically we will need 512 steps or 9 bits scanning to achieve +-0.01D mathematical accuracy due to round-off errors and neglecting other factors. Question: what if flatbed scanner is good quality and scanning is done in 16 bits? Wouldn't it be sufficient?

    For UV and polarizing filter cannot we improvise by putting something on the scanner glass?

    Quote Originally Posted by spider View Post
    because your scanner doesnt have a standard d50/d65 light source.

    the spectrum of your scanner light is not defined and the scanner light has not exactly the same intensity over the whole scanning area.

    also the degree of the light source and the ccd is not the same as for spectrophotometers.

    non existing uv-filter might be another problem.
    Last edited by Gargoyle; 11-12-2012 at 03:03 AM.

+ Reply to Thread

Similar Threads

  1. Scanner
    By huanun in forum Graphic Design
    Replies: 11
    Last Post: 02-08-2012, 03:57 AM
  2. Buying ED 100 efi densitometer
    By sanimir in forum Requests
    Replies: 0
    Last Post: 12-30-2011, 07:12 AM
  3. Flatbed printer kit
    By metttiu in forum Requests
    Replies: 0
    Last Post: 04-22-2011, 02:51 AM

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