Rik I am debugging as you suggested, and I found a mismatch between the formula and Photoshop's results. A is top image and B is bottom image in the attachment. The blend mode I apply in Photoshop is Multiply, just to keep calculations easy. R and B components match, but for G I get different result with the formula: G = A*alpha_A*(1-alpha_B) + B*alpha_B*(1-alpha_A) + alpha_A*alpha_B*Multiply(A,B) = 0.75*0.75*0 + 0.75*1*0.25 + 0.75*0.75*0.75= 0.609375 and that is different than 157 for G in Photoshop which maps to 157/256=0.61328125 in range 0 to 1. So in summary, I get same results for R and B, and for alpha, but different for G.
Received on Friday, 27 July 2012 22:33:59 UTC