Digital Light & Color

Except when the brightness curve lies above the diagonal, adding a new control point (by shift click) straight on the smooth curve flattens partly the curve in both directions from the new control point influencing contrast in the corresponding parts of the image, which requires, if unwanted, additional control points to compensate for it even before the new control point is moved anywhere. I would expect no change to the smooth curve until the new control point is dragged. Has it always behaved like this and I have just noticed it only now?
If this is the intended behavior, what is the advantage over keeping the smooth curve unchanged until the new control point is dragged to a new position? Keeping the curve unchanged would enable anchoring its other parts with control points before the new point is moved elsewhere. Now anchoring the smooth curve to freeze parts of its shape is not possible because it changes its shape with every new added control point. The same is true with the color curves.

David

PWP has worked this way since the beginning. There is an extensive body of mathematics behind passing curves through a set of control points (called knots) based on splines. The family of spline curves Photoshop uses can overshoot producing unwanted solarization, so PWP is a big advance in that area. While the Carlson-Frisch monotone splines PWP uses never solarize unless the control points require it, they do not have the property that adding a point on an existing curve preserves the slope at that point, so you may have to add more points to get the effect you want. I am not aware of any monotone-preserving spline that also preserves local slope when you add a new control point.

Jonathan,
thanks for the explanation. I am sure I can go on with PWP curves quite happily, the more so I can only very remotely guess what is going behind the scenes. :)
By the way, this gives me the necessary peace of mind and no need to compare with other tools I do not use nor need.

David