Understanding Blend Modes: the DIFFERENCE blend mode 

Really awesome, you explain it so clearly and with so many practical tips, thank you very much.

Great. Thank you for another great installment. I had to really study the visual effect of the blended B&W gradients to understand. I guess the reason for the black diagonal line in the result image is because the arithmetic difference between the pixels in the blend and base layers along that diagonal path is zero (0-0=0; 128-128=0; 255-255=0). And the reason that the upper left and lower right corners are white is because the arithmetic difference between the pixels in the blend and base layers is maximal (abs(255-0)=255; abs(0-255)=255). As for the practical applications, the alignment check is the easiest to comprehend. The next example works to darken a base layer image by shaving a few points off the RGB values of that base layer. So, a base layer color that is r150, g200, b60, when Difference blended with a dark grey layer that is r10, g10, b10, becomes r140, g190, b50. Is that correct? BUT, if your blend layer strays too far from black, the absolute difference between blend and base layer pixel may yield a bigger value, inverting the values of the original base layer. Is that correct? Related, if your blend layer color moves from black toward red, the base layer red values drop initially revealing a complementary color cast until the absolute difference reverses the trend and restores base layer red values. Is that correct?

Seriously, you and Dave Straker of InAffinity should get together for dinner. I would imagine the two of you would have a fascinating conversation regarding Affinity Photo. 🙂