Often, defining either positive or negative control points, a circle is not always the best shape. For example, if I want to affect a rectangular building, a circular positive mask often means I have to apply lots of negative masks to limit the effect to the rectangle. The same principle applies to negative masks, where they can often overlap a positive area by too much or need quite a lot of negative masks to effectively mask a rectangular positive area.
My suggestion is that we should be able to choose the shape of the control point mask.
It would certainly be useful but could be a UX challenge. In drawing apps, ellipses are drawn by dragging the corner of a bounding rectangle, where a control point is manipulated by grabbing the (circle) shape itself. How does one drag the edge of an ellipse in a sensible fashion?
My concern about this proposal, Joanna, is that it would probably reinforce the oft seen misunderstanding that the shape of a U-point’s “area of influence” somehow determines the area selected by the U-point … esp. for new users coming from other applications.
And the fact this this has received 15 votes (at time of writing) suggests to me that this is already the case !
I agree with your thoughts on being misleading. Would you believe it’s taken me a couple of years to finally get a grasp on how best to use control points effectively? See my post here
In fact, as powerful as the control point tool is, it’s phenomenal capabilities are very inadequately covered in the manual and other tutorials.
The impression you get from what is out there is that it just needs one (maybe a couple) point to cover an area, surrounded by loads of negative points to constrain the effect.
In my recent experience, it can actually take several points to include differing tonalities in a desired area, otherwise parts of the area that differ in tonality do not get effectively selected.
So, in addition to raising the subject of changing the shape of the affected area, might I also raise the subject of better documentation of what it does so far?
Control points are a very powerful tool.
The problem in PhotoLab is poor implementation: The worst of all known software with this technology.
Last week I installed Snapseed on my smartphone. Even in this free app, the selection of control points is much better.
I continue to argue that the implementation of control points in a PhotoLab should be fundamentally rewritten. Until that happens, I will continue to use this technology in Capture NX2.
In the topic for the most desired updates in the PhotoLab, I already wrote that this is one of the most desired improvements in the PhotoLab 5. This will be one of my arguments whether to update to PhotoLab 5.
P.S .: I have been working with control points for more than 10 years, so I think I know the technology well enough …
I don’t doubt that’s the case for you, Sigi … but I’ve seen a lot of posts in the time since Control-Point technology was introduced in PL where it’s obvious that it’s not understood. (And I remember this being a learning process for me too - in the early days of PL’s Local Adjustments).
Yes, that’s my experience too - but I find U-points quite easy/simple to use when “M” (for mask) is applied to reveal the actual area(s) of selection - and helping show to where negative U-points are needed.
Absolutely - - I agree with you wholeheartedly on this need.
I agree with Joanna. Sometimes another shape is better.
Standard shape: Circle.
Alternative shapes: Rectangular and elliptical.
Stellar option: Shape drawn with free hand.
Were it possible to copy a point alternatively including original color or to be able to expand the affected area with a brush once the point has been established?
Personally, I feel the circle is good for most jobs but occasionally another shape might cut down the working time.
Maybe I am exactly the type of person you are talking about but I do not agree with that statement.
In the below, extreme, example the affected areas are very obviously circles. Because the strength of the effect varies by linear distance from the control point, the circle shape has a significant effect on the result.
I know that the effect is tone-specific, but even then, areas of the selected tone that are near the edge of the circle are less affected than those closer to the centre. With an elliptical control point, the strength would be increased further in absolute terms along the major axis than the minor axis.
Furthermore, sometimes one wants the effect to avoid an area of the photo that has the same tone and requires negative points to achieve this. Having a more malleable area of influence would reduce the need for negative points in some cases.
The single biggest improvement to understanding of control points, in my mind, would be to somehow promote the “show masks” checkbox so people will naturally want to click it.
Actually it would be useful to have Control Point tech without masking to apply vignette or lighting effects. I’m a fairly new user so apologies if there already is such a thing!
Duh - caught out! A feeling stupid moment. I was thinking something like the lighting filter or gradient filter in Photoshop. On CS5 both are fairly basic.
I agree wholeheartedly. I also still use the control point technology in Capture NX2, which i find much more intuitive in defining the area of impact compared to the control point implementation in DXO.
I use to be very frustrated by the strength of the effect dropping off with distance. I often had far more selection points than I wanted and found that I was getting spotting that was unacceptable. I eventually realized that in most instances if I just make the circles MUCH bigger it solved the problem. In this example it was just a few clicks to isolate the plane from the blue sky.
) !
