1 Reply Latest reply: Feb 4, 2010 5:31 PM by George Austin RSS

    Missing One of Three Blend Mode Constraints

    George Austin Community Member

      It takes three independent parameter constraints to fix the three-channel color value result of a blend. For each of the four related blend modes Hue/Saturation/Luminosity/Color it is readily seen that hue and luminosity are precisely fixed as either the bottom or top layer value, depending on the blend concerned. That gives us two constraints. What is the third ?

       

      I can tell you what the third constraint is NOT for ANY of the four modes: It is not the saturation of the top or bottom layer---a tempting guess because it would give parametric symmetry for hue, saturation, and luminosity or because that is asserted (incorrectly) by PS Help. Nor is it the brightness value of top or bottom layers.

       

      Some have suggested that the issue might be resolved if the definition (Max - Min)/(Max + Min) were used to define saturation in lieu of the more common (Max -Min)/Max, thinking that perhaps this alternative saturation of top or bottom layers might be preserved. But neither top nor bottom saturation by either definition is preserved.

       

      Interestingly, the hue and color modes both use top hue and bottom luminosity, and so would be identical were it not for the the third constraint which has to differ for these two modes because they do, in fact, produce different resulting colors.

       

      So, we can nail down resulting hue and saturation for these blends, but we're missing the last piece of the puzzle. Close but no cigar! Repeating the question, what is the third constraint for each of the four blend modes cited or, for that matter, the third constraint for any one of those blends?

        • 1. Re: Missing One of Three Blend Mode Constraints
          George Austin Community Member
          That there have been no responses may be due to my poor presentation of the question. I'll  try again.
          PS books and other sources I have checked give the parameter values for hue, saturation, and luminosity resulting from Hue, Saturation, Luminosity, and Color Blend modes as  either the base or blend value, depending on the mode. While this is correct for hue and luminosity, saturation values resulting from any of these four blends DO NOT comply. The definitions used by the authors are supported by Photoshop Help, and probably have been  engendered by PS Help. In practice, however, they do not pan out for the saturation parameter. Let me demo the saturation deviation by just one example, taken at random from more systematic sampling (saturation values are in red):
          Base color: RGB 177/25/37, HSB 355/86/69, luminosity 72
          Blend color: RGB 115/74/54, HSB 20/53/45, luminosity 84
          Resulting colors:
          Hue Blend Mode: RGB 146/49/1, HSB 20/99/57, luminosity 72
          Sat Blend Mode: RGB 114/53/58, HSB 355/54/45, luminosity 72
          Lum Blend Mode: RGB 189/37/49, HSB 355/80/74, luminosity 84
          Col Blend Mode: RGB 103/62/42, HSB 20/59/40, luminosity 72
          NONE of the resulting saturation values match either the base or the blend saturation values, whereas ALL of the hue and luminosity values do. Luminosity values are given by 0.3R + 0.59G + 0.11B, the definition used in PS, but their tedious calculation can be  skipped via a neutral layer in any of the four modes placed above the blend layer and turned on only to read luminosity (the resulting gray value).
          This leaves a problem. While luminosity and hue constraints on the resulting color are exactly defined, we still need a third constraint in order to fix color---which requires three independent parameters.  Brightness of the base or of the blend is NOT preserved. If it's neither brightness nor saturation, what is that third constraint?
          EDIT: Hard returns have been ignored. Even multiple hard returns do not separate paragraphs as before. So please excuse the crammed format.