U-point selection centre area. just thinking

U-point is color selective in its masking.
i don’t know how precize, small, the selection area is in side that cirkel you use to select the influence area. but if you want mask a color spotted area like tree leaves, brikwall, grasland and such, it could be difficult to choose the right spot for the right influence area.
Would it be usefull to make the selection centre variable in size to make it easier masking spotted area’s?
like a cloningtool.
selecting the cirkel radius which is used to paste. So you can select a colorpalette of the area you want effect with a u-Point. This would be working faster then using brushes and removaltool to select, mask a object or area like tree branches

(i could be wrong in my understanding of the tool so maybe its abogus idea.)

In this case, the diameter of the first disc needs to be small.


Hello Pascal,
Can you elaborate some more? Are you mean to make the first control point small in circle and copy those after that in bigger upoint circles in the chosen area?

As far as i know is the pinpoint of the Upoint also a point where a cluster of pixels is examend for determine the mask (which pixels are effected by the upoint). The circle around this pinpoint you can alter in size is only a way to spread the influence section of the mask.
What i mean is would it be interesting to control that pinpoint diameter and thus the cluster of pixels. And thus the amount of colors inside this group/cluster of pixels
This way you can select more or less pixels to control the chosen color(s) in a area.

1 Like

Sorry OXIDant (it is your name ?)

Wrong answer based on my first understanding of the U-points when i discovered the functionality.
No, each circle is different and have its own influence area. It takes into account density - brightness and color - of a group of pixels covered by its reticle to apply the correction only to the pixels that have the same characteristics within the selection disc.

You have to multiply the discs by putting them well on the leaves of the same colors.

No its Peter. :slight_smile:

Thats that three peaces of a circle in blue is suppose? translation is crosshairs.
So if a section has three colors next to each other i have to give each color a upoint selection which cirlces overlapping to influence all three colors the same.
(you did a good job in this leaves changing color btw)

Those upoints are much more powerfull then the tutorialsection of DxO showes at this time.

What i would be able to see is which pixels (color) is picked by the reticle if i place it somewhere. So i can place it more accurate.
One thing i notice now is that upoints in nik and PL are a great tool to handpick sections in all kind of situations.

I found the word “reticle” the good name to describe the task.

If it is to influence 3 colored areas in the same way, there could be better tools than the U-point !?

That is the condition.
That’s why I think U-point is not a versatil tool.
This example (plants leaves) shows a specific use where that it functions very well.


Ok got in front of the screen testing color behaviour.
Using the default “hand” the squarebox next to the sun of the histogram is showing the actual color be neat the “hand” (it would be great if it also shows a croshair locating the actual place in the histogramfigure but that’s another feature i would like to have. (now it only shows color(squarebox) and RGB numbers and if you like to find which part of the image causes the spike in a hystogram that croshair system would help a lot.)
Activating local adjustments controlpoint:
The “reticle” has a “eye” in the centre which shows the color behind the controlpoint:
But the squarebox in histogram is gone:
Turning off local adjustments, place the hand on the same spot as the control point reveals the color in the box which is the same as the color in the “eye” of the controlpoint/reticle.

To put the “eye” which reveals the color/and luminance (and all other characteristics it influences by the controlpoint) more precize in your image you could use the wheel on the mouse to zoom in to 350-400% and place it on a pixel of desire(notice that the reticle isn’t resizing wile zooming so zooming out wil cover again a larger cluster of pixels.) but i would be liked that the histogram’s info is linked to that reticle if selected.
So you see:

  • the color which is inside the “eye” in that square box.
  • the RGB numbers of the color (incase you write it down you can find the same color on other places of the image to put a second controlpoint)
  • the crosshair in hystogram shows the location of the “eye” in the histogram figure for sake of luminance spikes or color spikes.

Second the three (blue) parts of the outer circle of the reticle
Represents as i suspect the area which is also used to determine the mask selection.
Or its just a symbol and has no meaning. But testing this eye in mask modes (m) holding down left mouse button to move the “eye” around you see a second grey circle with is fit inside the blue circle. and moveing it around on a hard color edge you see the mask changing when the grey circle crosses the edge wile the “eye” is still completely inside the hard color.

So it lookes like the area between the “eye” and the blue cirle is used for “feathering” purposes.
So if i could resize that space between the two circles i would influence the “feathering” of the mask. (usefull when a color isn’t equal but spotted/flaked like leaves or stone or such to exclude or include pixels around that selected color in the “eye”.
(like the controls you have in masking brushes:

I don’t know the algorithm/coding behind controlpoints so i don’t know if its possible at all. but i think it will add a lot more versatile use for it.

1 Like

Thinking this through this could be very difficult:
This point in the cros of the croshair is a sum of 4 channels : luminance, Red Green Blue representing a 3D point in a colorspace. So can that be marked in a 2D histogram?

YES Peter, this seems to me an exhaustive description.

That’s exactly what I wanted to ask without thinking about the shape of the tool.
Hope that DxO staff read your comment.

About the histogram crosshair, this is a old demand. I don’t know if it’s plan by DxO.


i think there bussy launching PL2.0. :wink:

Nevertheless i would see more detailed manual/tutorial/webinair of blending in local methodes as gradient filter/(auto)masking and upoint controlpoint technology in the basis workflow of images. If they extend the knowledge of the now availble toolpower it would be helpfull for the “weakpoints” some users encounter using PL.
Because i am convinced that Ad**e 's extended and overwhelming infomation how to’s more the problem is then that it is a better tool for users to switch over to DxO’s system.
But that’s just my opinion. :slightly_smiling_face: