Loading ...
Sorry, an error occurred while loading the content.

WB-blending with Tufuse

Expand Messages
  • Flemming V. Larsen
    I often have to take photos under difficult lightning and WB conditions fx. mixed interior- and daylight. Normally I d extract two pictures from the raw file
    Message 1 of 5 , Apr 6 2:32 AM
    • 0 Attachment
      I often have to take photos under difficult lightning and WB conditions fx.
      mixed interior- and daylight. Normally I'd extract two pictures from the raw
      file with different WB and blend them manually with masks in Photoshop.
      Often a rather time consuming task.
      So I've been on the lookout for a more automated way of doing this - I did
      try Enfuse a couple of month ago, but wasn't satified with the result: Still
      too much blue on the windows and too much red on the interior! When I got
      Tufuse I though it would give the same result, so at first I didn't consider
      it worth a try.
      But then I gave it a try afterall - and was surprised how well it was able
      to take the best from each WB into the fused picture :-))

      http://fvlmedia.dk/tokina107/wbtest/

      best regards

      Flemming
    • AYRTON
      How nice Thanks for sharing AYRTON On Sun, Apr 6, 2008 at 6:32 AM, Flemming V. Larsen ... -- ... + 55 21 9982 6313 http://ayrton360.com
      Message 2 of 5 , Apr 6 6:12 AM
      • 0 Attachment
        How nice
        Thanks for sharing

        AYRTON



        On Sun, Apr 6, 2008 at 6:32 AM, Flemming V. Larsen <fvl@...>
        wrote:

        > I often have to take photos under difficult lightning and WB conditions
        > fx.
        > mixed interior- and daylight. Normally I'd extract two pictures from the
        > raw
        > file with different WB and blend them manually with masks in Photoshop.
        > Often a rather time consuming task.
        > So I've been on the lookout for a more automated way of doing this - I did
        > try Enfuse a couple of month ago, but wasn't satified with the result:
        > Still
        > too much blue on the windows and too much red on the interior! When I got
        > Tufuse I though it would give the same result, so at first I didn't
        > consider
        > it worth a try.
        > But then I gave it a try afterall - and was surprised how well it was able
        > to take the best from each WB into the fused picture :-))
        >
        > http://fvlmedia.dk/tokina107/wbtest/
        >
        > best regards
        >
        > Flemming
        >
        >
        > ------------------------------------
        >
        > --
        >
        >
        >
        >


        --
        ------------
        | A Y R |
        | T O N |
        ------------

        + 55 21 9982 6313

        http://ayrton360.com
        http://rio.360cities.net
        http://vrfolio.com
        http://ayrton.com


        [Non-text portions of this message have been removed]
      • Erik Krause
        ... Nice test. This is interesting indeed. You didn t mention the weight parameter settings for enfuse so I tried with different settings. Moreover tufuse uses
        Message 3 of 5 , Apr 6 7:04 AM
        • 0 Attachment
          On Sunday, April 06, 2008 at 11:32, Flemming V. Larsen wrote:

          > So I've been on the lookout for a more automated way of doing this - I did
          > try Enfuse a couple of month ago, but wasn't satified with the result: Still
          > too much blue on the windows and too much red on the interior! When I got
          > Tufuse I though it would give the same result, so at first I didn't consider
          > it worth a try.
          > But then I gave it a try afterall - and was surprised how well it was able
          > to take the best from each WB into the fused picture :-))
          >
          > http://fvlmedia.dk/tokina107/wbtest/

          Nice test. This is interesting indeed. You didn't mention the weight
          parameter settings for enfuse so I tried with different settings.
          Moreover tufuse uses a different strategy if it encounters images
          without much exposure difference. It assumes these are shots from a
          focus stack and uses the contrast criterion only.

          But enfuse gives a different result even if the contrast criterion
          alone is used. This might be due to the fact, that tufuse uses the
          original approach to determine contrast edges (a laplacian operator)
          while enfuse uses a modified one (the local standard deviation).

          However, both applications are not intended for WB blending. It works
          only because (by chance) the proper WB has more local contrast.

          A better way to do WB blending is to use the contrast enhanced b
          channel of a Lab colospace copy of either image to create a blending
          mask (sometimes the blue channel would do as well).

          best regards
          Erik Krause
          http://www.erik-krause.de
        • Max Lyons
          I think that, by default, Enfuse uses a non-zero weight when evaluating the saturation of a pixel. And, by default, TuFuse uses a zero weight for
          Message 4 of 5 , Apr 6 7:09 AM
          • 0 Attachment
            I think that, by default, Enfuse uses a non-zero weight when evaluating the
            "saturation" of a pixel. And, by default, TuFuse uses a zero weight for
            saturation, caring only about exposure. This is the most likely explanation
            that comes to mind for explaining why the results look so different!

            Max

            > I often have to take photos under difficult lightning and WB conditions fx.
            > mixed interior- and daylight. Normally I'd extract two pictures from the raw
            > file with different WB and blend them manually with masks in Photoshop.
            > Often a rather time consuming task.
            > So I've been on the lookout for a more automated way of doing this - I did
            > try Enfuse a couple of month ago, but wasn't satified with the result: Still
            > too much blue on the windows and too much red on the interior! When I got
            > Tufuse I though it would give the same result, so at first I didn't consider
            > it worth a try.
            > But then I gave it a try afterall - and was surprised how well it was able
            > to take the best from each WB into the fused picture :-))
            >
            > http://fvlmedia.dk/tokina107/wbtest/
            >
            > best regards
            >
            > Flemming
          • Erik Krause
            ... In my tests mentioned in the other mail wSaturation was set to zero of course for enfuse. However, I don t see a good chance to do automatic WB blending.
            Message 5 of 5 , Apr 6 8:12 AM
            • 0 Attachment
              On Sunday, April 06, 2008 at 10:09, Max Lyons wrote:

              > I think that, by default, Enfuse uses a non-zero weight when evaluating the
              > "saturation" of a pixel.

              In my tests mentioned in the other mail wSaturation was set to zero
              of course for enfuse. However, I don't see a good chance to do
              automatic WB blending.

              Perhaps with an inverted saturation mask, favoring the pixels with
              less saturation. But this would probably give strange results in some
              cases: I imagine a light yellow object lit by outside light. With an
              "inside" tungsten WB this would appear light blue, with "outside"
              shadow WB it would appear bright yellow.

              The result would be light blue beause that's less saturated than
              bright yellow, probably more blue than the outside scene...

              best regards
              Erik Krause
              http://www.erik-krause.de
            Your message has been successfully submitted and would be delivered to recipients shortly.