Digital Light & Color

While there is an obvious issue of not overcomplicating this composite transformation, and one can always resort to using the full-version of Selective Colour instead, I would find it quite useful if the Show Affected Region check box could be included (there is a perfect room for it right next to Show Modified Colour Wheel...).

Also a question: when the mouse pointer hoovers on the colour wheel of Selective Colour, it changes to a cross at either end of the color mapping arrow. But when is over the arrow line, the mouse cursor changes to a white up-pointing arrow - what's does it signify? That the right-click options will pertain to this control point/arrow?

I'm not sure adding a Show Affected Region checkbox is a good idea, but I will think about it.

The up-arrow apparently indicates that, for example, clicking will select the control point and right clicking will apply to the control point, but dragging will not move either end.

I started using Adjust more often and I understand that its an amalgam of individual transformations with simplified controls, and if more controls are needed one can simply use the full version of the individual transformations, but I keep missing Show Affected Regions in Selective Colour tab of Adjust. I find it more useful in visualising what range of colours may be affected than Showing the Modified Colour Wheel.

Also, I'm a bit confused with the Weight slider. Help implies, as I understand it, that the Weight regulates the range of similar colours affected by a control point, but what I seem to be seeing, adjusting Weight doesn't change the range of affected colours, but rather regulates the intensity of changes caused by an arrow from a control point, if any, for the constant range of colours to be affected already defined.

For instance, if the control point is used to pin similar colours to limit the range of colours that another control point may be affecting (i.e. a control point with no arrow), moving the Weight slider doesn't change the range of colours affected by this control point nor does it change anything since the control point without an arrow pins the range of existing colours rather than changing them.

I added Show Affected Region for the next release.

I think the Weight slider is working the way it is described. The easiest way to see how it works is to check the Show Modified Color Wheel box. and experiment.

Thank you!

I guess what I have trouble understanding and differentiating is which range of hues/saturations will be affected by a control point placement vs. how much will they be affected by the control point destination colour and weight.

For instance, what does save as mask really represent - which is similar to Affected Region, sans colour changes? The way I understand it is that the mask brightness represents the degree that various hues/saturation that this control point can possibly change (black being beyond its reach). That doesn't change with the Weight slider - what changes is the intensity of the destination/target colour, which will only change according to mask brightness. Am I correct in thinking about it this way?

It's a little more complicated...

Each pixel in the output image is affected to some extent by all the control points (although that effect may be negligible). A color that falls exactly on a control point is always changed to its target color. Color and brightness changes are computed as a weighted average of all the control points where the weight falls off with distance from the control point and is also affected by the Weight slider. Increasing the weight of a control point thus causes it to more strongly affect a larger range of colors around it.

The way this transformation works is to build a full 2-D color wheel for every possible value, as represented by the Modified Color Wheel. Then each input pixel is mapped onto the color wheel and changed accordingly. The brightness is optionally adjusted by the Brightness Change sliders.

Thanks again!

So would it be correct to say then that the Save as Mask image (and the same implicit mask in the Affected Region) for a control point shows merely a map of 'colour distances' based on the difference in hue and saturation values between the exact colour of the control point and the colours of all other pixels in the image - brighter being more similar to the colour of the control point?

I just rechecked what the code does. The mask image for a given control point corresponds to the weight of that control point for a given hue and saturation as a proportion of the total of all the weights of all the control points. The mask will be white for any color whose hue and saturation exactly matches that of the control point.

I'm still developing an intuitive feel for the weight scale 0..100. It is easy to fall back to spraying fixed control points around, or, if only one color needs change, making a mask.

The way I would use it is to first set the control points and then adjust the weights afterwards if necessary to fine tune the results.