barninga

@jkehoe, the #7 shows a way to mask/unmask an effect in order to apply it only to certain areas of the image. in AP, any effect layer is also a mask. if you select an effect layer and paint black on it, the effect disappears where you just painted. if you paint white over the black area, the effect reappears. the exercise instruct to add a hsl layer, then invert it. inverting an effect layer actually means inverting its mask. when you add an effect layer, its mask is white by default: this means the effect is applied to the whole image. if you invert it, its mask becomes black and the effect is not applied anymore. in the exercise, you set the paintbrush to white, and the paint over a specified areas after selecting the hsl layer. this way, you create a white "hole" in its mask, which the effect becomes visible again through.

a dedicated graduated nd filter would be an interesting plus, ihmo. easy to use, but limited to its scope. the ability to create graduated filters of any kind is obviously more powerful. however, the fact that a gradient affects a mask destructively (as far as i could understand by experimenting) and you need to use a vector shape as a clipping mask to apply a non destructive gradient sounds like a bug to me, or at least like a flaw in AP's mask implementation (it's not the only one; i'd add the inability to dodge and burn in masks).

i never used that program, so i cannot make comparisons. however i used other programs and at the beginning i found a bit hard to learn AP's logic. now i'm happy with it.

@Forssux, you can apply a gradient mask to adjustment layer: just use the gradient tool on the adjustment layer. afaik, this technique will blend the gradient into the mask permanently. if you'd like to have a non-destructive gradient, you can use a shape (usually a rectangle) as a vector mask: apply a gradient to the shape and set one of its endpoints' opacity to 0.

@Sima: yep, it's the same thing - even if in your test the artifacts look smoother. @v_kyr: yes, i mean that. so it looks like something goes wrong, maybe in the process of creating the new layer or just in the rendering step. i'll wait a little to see if someone from the AP crew reads these posts and gives us an explanation; if not, i'm going to repost this in the bugs section of the forums. thank you for your replies and your time.

hello all, and happy new year! please i need some explanation about refining selections to "new layer and mask". i did the following experiment on the attached image: 1) select the background of the image with the selection brush 2) invert pixel selection 3) refine selection (default values, except for a bigger brush) 4) select "new layer and mask" 5) apply i'm not interested in how the quality of the selection could be enhanced (i.e. applying some feather during the refine process): instead, i can't understand how the new layer (with mask) is generated. I would expect a copy of the original layer, with a mask that describes the selection. the mask appears how expected (by me at least), but the new pixel layer shows something weird (to me). in the attached image, if you hide everything except "layer - created by new layer and mask", strange artifacts are clearly visible. they disappear if the mask is activated again. if i make a copy of the original layer and apply a copy of the mask to it, the result is similar - but not identical: if you alternatively make visible the new layer or the background copy (with their masks), you can notice that something in the image subtly changes. if you make both the layer visible at the same time, the quality of the selection improves. it improves even more if you make a second copy of the background copy and mask. now, the improvement of quality does not surprise me so much: i guess it depends on the fact that the areas that are left erroneously slightly transparent by the mask get more opaque when displayed one above the other. but i really cannot understand why the new layer created by "refine selection" is not a copy of the original layer, and what the artifacts are. i'll be grateful of an explanation or any clue. thanks in advance stefano IMG_1904.afphoto

@RNKLN, thanks for sharing.

i cannot reproduce the behavior you described. be sure you select the layer -click on it in the layer stack- where you want to use any tool before you begin.

yes @R C-R you are right when referring to adjustment layers as we know them in AP (and in PS, and other photo retouching programs, i suppose). However, we could think about a special kind of adjustment layer, where the adjustment is not globally applied, but only locally, where painted by a healing tool, or some other kind of brush. A sort of mix between actual adjustment layers and regular pixel layers. it could implement several kinds of "relativeness" to the underlying layer - hue, luminosity, and so on. i don't know if it's possible from programming point of view, and how much it could be tricky.

@R C-R i think @N'Awlins Contrarian means something like this: when the healing tool is applied to copy the pixels from an area to another area on another layer, what should get stored with this second layer are not the actual pixels, but the difference in hue and lightness between the pixels that are copied and the pixels that they cover ("heal"). at least, i think.

great suggestion @Gnobelix! i missed the batch item in ap menus. thank you!

well i use an imac running sierra and i do run bettertouchtool also (great little piece of software!) and i have no issues in resizing brushes with ctl+opt+drag. double check the mouse and trackpad settings in system preferences, there could be some conflict (i vaguely remember i had to check or uncheck something somewhere to make things work as expected). about [ and ], i think i had your same issue (i use an italian keyboard) and solved it by setting "è" and "+" as brush resize keys in AP preferencecs (è and + are the characters you get on an italian keyboard by pressing [ and ]; to actually get the square brackets you have to press the right opt key too -- it looks like AP checks the character code rather than (or in addition to) the key code to determine what key was pressed. keyboard shortcuts are highly configurable in AP, maybe with some experimenting you'll be able to get the uniformity of configuration you spoke about.

yep. thanks for pointing that out. apparently they could not fix for the 1.6 release yet.

thank you for your replies. yes i think there is something wrong in the way ap handles luts, and unfortunately i don't have any other application lut-aware. since, in my turn, i use to copy adjustment layers and paste them to the images where i need more or less the same adjustments, luts are a low priority item for me too. i was just experimenting, since from @Mr. K's reply i caught that something could be wrong. it looks like it is, actually.

yes @v_kyr, your explanation is clear and sounds good to me because you effectively described more or less what i had understood of the whole matter. thank you. what i can't understand now, is the apparently strange and incoherent results i obtain by doing very simple operations of lut exporting and importing. basically, if i have 2 identical images, alter the colors of the first, export its lut and import it into the second image, i'd expect to obtain an image the presents (approximatively, at least) the same color alterations i did in the first one. if you look at the pictures i attached to my posts above, you can get what i mean.