Selection tools seem to work on image layers, when in fact they do not

[Frank M.]

Problem found in Windows 10 Home 64 Bit, Affinity Photo 1.3.3 (481).

When starting a new photo using the menu point "File" -> "New from Clipboard" (or "File" -> "Place"). The layer with the image is of the type "Image" rather than "Pixel". Apparently, one can select using any of the selection tools, but when one tries to copy the selection to a new layer (ctrl-j) or the clipboard (ctrl-j), the whole layer gets copied rather than just the selected areas. If the select tools are not to work on "Image" layers, they should not appear to be working, i.e. one should not be able to mark areas with running borders around them with the marquee tools, flood select or paintbrush select. Otherwise, one may get very confused as to what is wrong. It had me baffled for some time until I noticed the layer type "Image", which I could then right-click and "Rasterize". 

The best UX would actually be dialogue box warning that selections cannot be done on image layers, or perhaps the new layer could just be rasterized to type "Pixel" right away.

 

Best regards,

Frank

[walt.farrell]

8 hours ago, Frank M. said:

Apparently, one can select using any of the selection tools, but when one tries to copy the selection to a new layer (ctrl-j) or the clipboard (ctrl-j), the whole layer gets copied rather than just the selected areas. If the select tools are not to work on "Image" layers, they should not appear to be working,

Welcome to the Serif Affinity forums, Frank.

If you're referring to the Marquee Selection tools or the Freehand selection tool, those do not work "on layers" at all. They make a free-standing selection area that will apply to whatever pixel layer happens to be active when some subsequent action occurs.

For example, if you had 3 pixel layers, you could have the top one active (selected in the Layers panel), draw the marquee, then switch to the bottom one do a copy. That would copy from the bottom layer, not the layer that was active when you created the marquee.

You can even create such selections when there are no pixel layers. They won't have any effect on actions until there is an active pixel layer, but they can exist.

That's the source of your confusion with the (Image) layers, I think, and it has confused other users, too. It's not really a bug, though, just a common misunderstanding of how the program works. But you're right that it should be improved in some way. Possibly there could be an Assistant setting to automatically rasterize the active layer, or to do nothing at all but warn if there's a pixel selection active when an action is taken that would act differently on the active vector layer.

[carl123]

9 hours ago, Frank M. said:

i.e. one should not be able to mark areas with running borders around them with the marquee tools, flood select or paintbrush select.

The best UX would actually be dialogue box warning that selections cannot be done on image layers, or perhaps the new layer could just be rasterized to type "Pixel" right away.

Selections on image layers are perfectly valid

Try making a selection on an image layer then adding a Live Filter such as a Gaussian Blur or add an Adjustment such as Curves or Recolour

[Frank M.]

Hi Walt,

Thanks for the clarification!

You are right in regards to what confused me. In particular, the fact that one can use the image layer as a basis for the paintbrush tool, i.e. the areas get selected based on the colours present in the image layer, although the selection cannot work on that very same layer. Even if selected areas exist separate from layers, it does seem very counterintuitive that one may select an area based on an image layer, but then one cannot apply move the selection to a new layer for manipulating it.

I will never make that mistake again, but for other, new users it would be much better if the "New from Clipboard" function just created a pixel layer right away, possibly with a setting to create an image layer, if one really wants that.

/Frank

[Frank M.]

@carl123

Yes, I have now tried that out. One can go directly to adjustments and any adjustments are implemented in a new adjustment layer affecting just the selection. One can even just create a new empty pixel layer on top (having the selection active) and then do e.g. a flood fill on the new pixel layer and only the selected area will be filled. 

However, I feel the above behaviour for a selection done based on an image layer actually gives even more weight to my argument. If I can do adjustments or changes on the selection which goes to the new layer (either automatically or by creating a new, empty pixel layer first to work on that), it would seem logically consistent that "Layer" -> "Duplicate Selection" (Ctrl-J), should also duplicate exactly my "selection" from the image layer and put that very selection onto a new pixel layer, rather than create a new, additional image layer with the whole image and thus ignore my selection. 

The function is called "Duplicate Selection", and I have a "selection", so please duplicate my "selection", not the whole image.

I am sure after one gets used to it, the behaviour will be internalized and justified -- In a few months time, I will most likely have lost the fresh perspective of a new user and find myself berating and educating the new, inexperienced users on why selection works just the way it should on image layers :-). For now, however, my viewpoint is that this behaviour/wording is inconsistent and simply confusing and therefore should be corrected, even if not a bug that crashes the application or spoils the data in an image.

/Frank

[walt.farrell]

55 minutes ago, Frank M. said:

The function is called "Duplicate Selection", and I have a "selection", so please duplicate my "selection", not the whole image.

Yes, but  you actually have two selections:

1. The vector layer, as an Image layer is a kind of vector layer; and

2. The pixel selection.

Actions that you perform are sensitive to the context in which you perform them. The "duplicate selection" action will copy the selected portion of a pixel layer, or the complete vector layer.

[Frank M.]

@walt.farrell

I admit that I am not so far with Affinity Photo yet that I understand the workings of vector layers. Even so, the concept that the image layer is a "kind" of vector layer and on top of that by default this whole layer is selected, although I have done nothing actively to select it, seems a bit convoluted and counterintuitive. I can't see why anyone would benefit from such default functionality.  It would make more sense to me if there was a pop-up message "No selection available" when I tried to use the "Duplicate Selection" function or even better, that the application just got the pixels of the selection behind the scenes (i.e. by rasterizing the image layer in the background) and created the new pixel layer with those.

Also, if the selection to be duplicated is the whole "vector" layer, should there not first be an active act of selecting it?

I am also trying to understand the benefit of getting an image layer when opening an image from the clipboard. Why is that different and better as compared to when I open a PNG file, which is then created as a pixel layer?

Thanks for taking the time to explain the functionality to me. Perhaps it is just parts of the workings of Affinity I do not fully understand yet, but do leave the door open for the possibility that I might actually be correct: That although the behaviour can be explained in some way by some kind of logic, that it may in fact not be very useful, sensible or intuitive as compared to just have the selection work the way one might expect.

 

/Frank

[walt.farrell]

57 minutes ago, Frank M. said:

Also, if the selection to be duplicated is the whole "vector" layer, should there not first be an active act of selecting it?

There was an active act; the layer was selected in the Layers panel, though that may have happened automatically when you created the layer rather than manually by clicking on it.

59 minutes ago, Frank M. said:

I am also trying to understand the benefit of getting an image layer when opening an image from the clipboard. Why is that different and better as compared to when I open a PNG file, which is then created as a pixel layer?

The benefits of Image layers are covered in the Help, and elsewhere here in the forums. But I am not sure why for the specific case of creating a new document from the clipboard it's better to end up with an Image layer. Sorry.

 

1 hour ago, Frank M. said:

Perhaps it is just parts of the workings of Affinity I do not fully understand yet,

I believe that is correct; you just do not yet understand all the possibilities and capabilities of the program. But there are aspects of Image layers, and of Marquee selections, that have confused enough new users that there is a good case for doing something to improve the usability and minimize the surprises. Perhaps Serif will come up with something in that area in the future.

[Frank M.]

@walt.farrell

9 hours ago, walt.farrell said:

you just do not yet understand all the possibilities and capabilities of the program.

Sure, and I look forward to trying everything out!
So far, I only have praise for Affinity Photo. It has been a real pleasure to work with it.

 

 

 

 

 

9 hours ago, walt.farrell said:

But there are aspects of Image layers, and of Marquee selections, that have confused enough new users that there is a good case for doing something to improve the usability and minimize the surprises. Perhaps Serif will come up with something in that area in the future.

Agreed!

 

Thanks, for welcoming me to the forum, Walt 

 

/Frank

[walt.farrell]

You're welcome, Frank. Good to have you here

[R C-R]

20 hours ago, Frank M. said:

I am also trying to understand the benefit of getting an image layer when opening an image from the clipboard. Why is that different and better as compared to when I open a PNG file, which is then created as a pixel layer?

AFAIK, New From Clipboard does not create an "(Image)" layer unless what is copied to the clipboard is itself an "(Image)" layer copied from some Affinity document. This is because only the Affinity apps (all 3 of them) are capable of interpreting the format of an "(Image)" layer as an "(Image)" layer.

If anything else is on the clipboard, it either will be pasted into an Affinity document in some format the app supports, or if it is not a supported type it generates an error message to that effect & nothing will be pasted.

This may be easier to understand if you also understand that the clipboard can simultaneously contain several different versions of whatever you paste into it. Assisted by the OS, the app tries to paste the most usable version of it into its document. Likewise, it also depends on what the 'donor' app passes to the clipboard, if it is a 'private' clipboard, & so on. (This is true for all apps that support the system clipboard, not just the Affinity ones.)

So for example, styled text copied to the clipboard might be pasted into an app while preserving all of its style attributes, effectively as RTF (rich text format) data, or as plain (un-styled) text, or with only some style attributes preserved, all depending on what the app can support.

All this may seem very complicated (because it is) but the most important thing to remember is you need to be aware of the layer type added to your Affinity document because that determines what you can do with it.

[Frank M.]

@R C-R

 

 

 

1

1 hour ago, R C-R said:

AFAIK, New From Clipboard does not create an "(Image)" layer unless what is copied to the clipboard is itself an "(Image)" layer copied from some Affinity document. This is because only the Affinity apps (all 3 of them) are capable of interpreting the format of an "(Image)" layer as an "(Image)" layer.

I copied the image from a Google Image search, right-clicking on the selected image, selecting the menu-point "Copy Image", and then used the "New from Clipboard" function in Affinity Photo, which gave an "Image" layer. So, it is apparently not only image layers from other Affinity Photo documents that can have this "Image" format. I thought the Google "Copy Image" function would just put the pixels of the image in the clipboard but apparently not.

If instead of creating a new document from the clipboard in Affinity Photo, one pastes the clipboard into a Windows Paint document, then in Paint press Ctrl-A and Ctrl-C to select and copy the whole image, and finally open a new document from the clipboard, A group layer is created with a sub-layer of the type "Pixel".

Thanks for the detailed overview of how the clipboard works!

 

/Frank

[R C-R]

3 hours ago, Frank M. said:

I copied the image from a Google Image search, right-clicking on the selected image, selecting the menu-point "Copy Image", and then used the "New from Clipboard" function in Affinity Photo, which gave an "Image" layer. So, it is apparently not only image layers from other Affinity Photo documents that can have this "Image" format.

You are 100% right about that. I don't know what I was thinking when I said otherwise. 

[Frank M.]

@R C-R

6 hours ago, R C-R said:

I don't know what I was thinking when I said otherwise.

Never mind that 

Your detailed analysis provided valuable insight into the actual workings of the "New from Clipboard" function and the fact that a copy Google images are not just a pixel selection, which was one of the causes of my problem and confusion when working on the initial image.

I still think, though, that Serif should change the "Layer" -> "Duplicate Selection" function from copying a whole "Image" layer that apparently is "selected" in an implied sort of way as opposed to the explicit and visible selection one is making when using any of the selection tools. As Walt mentioned, many new users are certainly going to be tripped up by this. Thus, I think that the "Duplicate Selection" function should just apply the active, visible selection to the image (rasterized in the background if need be) to create a new pixel layer with that selection. This will be more intuitive and better UX, in my opinion.

If one really just wants to copy the whole, "selected" layer, there is a logical and more evident "Layer" -> "Duplicate" function. This is also why I find that for the "Duplicate Selection" function to prefer to go by another, lesser (semantically) meaning of the word "selection" is confusing. I am certain that many users when seeing an active border around an area selected will instinctively see that as the selection to be worked with or upon.

While I may not be that experienced with Affinity Photo yet, I do have many years of working experience in software and user interface design.
I may, of course, change opinion later on and if so, I will come back here to readily admit it 

All of this is a bit nitpicking, which I also like.
All in all, though, I must say that Affinity Photo is a great tool that just grows on you as you discover and learn all of its efficient features.

A nice weekend to all of you!

 

/Frank

[R C-R]

6 hours ago, Frank M. said:

Thus, I think that the "Duplicate Selection" function should just apply the active, visible selection to the image (rasterized in the background if need be) to create a new pixel layer with that selection. This will be more intuitive and better UX, in my opinion.

(Image) layers retain their own pixel resolutions independently of the document they are in. So for example, an .afphoto document with a canvas size of 800x600 px & a document resolution of 96 DPI could have a 3264x2448 px photo from an iPhone or whatever placed or pasted into it as an (Image) layer.

To exactly fill the canvas this layer could be scaled down to ~25% of its 'native' pixel size, so in the document it is 3264x2448 px @ 391 DPI. (If you were to select it with the Move Tool, the context toolbar would show you that.) Of course, it also could be resized to other scales & DPI's, including ones with different horizontal & vertical scales & DPI's.

Because a 'marching ants' selection just defines an area of the canvas, not what is in any of its layers, if an (Image) layer is the current document selection -- note that these are two different kinds of selections! -- what exactly do you think should be rasterized in the background to create the new (Pixel) layer?

[Frank M.]

16 hours ago, R C-R said:

an (Image) layer is the current document selection -- note that these are two different kinds of selections!

From a UX perspective, this is exactly why the confusion arises for many (new) users, i.e the function "Duplicate Selection" most of the time means the area with a running border on the canvas as applied to a pixel layer. Even when one has a pixel layer in the photo document, this will also be the "document selection". But no one expects Affinity Photo, in this case, to suddenly decide to use the "document selection" as the selection to be duplicated. It is obvious for everyone that it must be the area selected on the canvas as applied to the pixel layer that is the "selection". Even if an image layer is the active "document selection", the choice of meaning of "selection" should not change. That is bad UX. A menu point should ideally be referring to just the same function or type of thing. In some cases, one may justify have two separate functions activated by the same menu point if it is extremely obvious which of the two functions is now activated, but it is usually not a good design choice.

So, it would be much better if "Duplicate Selection" always referred to the area of the canvas selected as applied to the active layer, if that makes sense. If not, the menu-point should just be greyed out or there could be a pop-up dialogue box saying "No selection active". If one wants to duplicate the whole image layer, there is a separate and logical function for that: "Layer" -> "Duplicate".

 

 

 

 

 

16 hours ago, R C-R said:

Because a 'marching ants' selection just defines an area of the canvas, not what is in any of its layers, if an (Image) layer is the current document selection -- note that these are two different kinds of selections! -- what exactly do you think should be rasterized in the background to create the new (Pixel) layer?

The most logical behaviour would be the same one that would happen if the image layer was just actively rasterized and so became a pixel layer. 

You mention that an image layer may hold an image of larger size than the canvas. This can also be the case for a pixel layer. I did a little experiment:

First I opened a new document with a small pixel layer. I then opened a larger JPG file in Windows Paint, used Ctrl-A and Ctrl-C to get that whole selection of pixels into the clipboard. I then pasted this into my AF document, and AF created a new image layer. Because the new image layer had a higher resolution than the canvas, only part of it was shown. Regardless, I could still use e.g. the Paintbrush Selection tool to select an area on the canvas. If I were to now use the "Duplicate Selection" function, I would want those pixels that I see as selected to go into a new pixel layer -- just the same way it would work if I first rasterized the image layer with the function "Layer" -> "Rasterize". This is basic rasterization that performs no other scaling or transformation. 

Of course, one might then realise that one would want to do some transformation first, which is perfectly fine -- the same would be the case if applied an area selection to a pixel layer of higher resolution than the canvas.

 

/Frank

[R C-R]

6 hours ago, Frank M. said:

So, it would be much better if "Duplicate Selection" always referred to the area of the canvas selected as applied to the active layer, if that makes sense. If not, the menu-point should just be greyed out or there could be a pop-up dialogue box saying "No selection active".

As has been mentioned, because a 'marching ants' selection just defines an area of the canvas, one can be made when there is no active (selected) layer, when more than one layer is selected, & even when the document is completely empty (as yet has no layers at all). There are workflows that rely on this (hint: for example, check out the options available in the Assistant Manager), so it would never make sense for "Duplicate Selection" to be greyed out (other than when no document is open, of course) because "Selection" in that menu item does not necessarily refer to any layer.

TBH, I am not sure why some users find the distinction between a 'marching ants' selection & any of the several other kinds of selections one can make in an Affinity document so hard to understand, or how that could be made any clearer than it already is. Consider that layers, text runs, vector nodes, foreground/primary & background/secondary colors, swatches, tools, & various toolbar items are among the many things that can be selected or active in the UI. Thus, "selection" or "active" or whatever else you want to call it always will refer to some specific context. I would think that much at least would be obvious within the first few minutes of trying out any of the Affinity (or many other) apps.

[Frank M.]

Hi "R C-R",

My real name is Frank.
What is yours?

I have been enjoying this exchange of viewpoints so far. It has uncovered some new information about the workings of image layers and Affinity in general, and we have stated our views. 

Unfortunately, I feel that your latest reply doesn't have so nice tone to it:

7 hours ago, R C-R said:

Thus, "selection" or "active" or whatever else you want to call it always will refer to some specific context. I would think that much at least would be obvious within the first few minutes of trying out any of the Affinity (or many other) apps.

There is no call for talking down to me as a new user. I just have a different perspective based on my current background and experience:

When I initially encountered the problem, I had been following an online course in Affinity for a few hours, including trying out some working examples, and I was now handling a photo to test the capabilities of the selection tools and adjustments. Prior to this, my experience had been Paint and Gimp in Windows now and then for just minor image touch up, i.e. no experience with any other type of layers or an application like Affinity Design. My knowledge in regards to photo handling was thus very likely similar to others that do not have deep knowledge of photo and design applications. In the first couple of hours of the course, the instructor had been covering pixel and correction layers as well as a lot about selections, but only selections in the context of areas of pixels selected in an image. This is also the meaning of the word "selection" that a user of primarily simple photo apps like Paint most likely would find relevant in relation to a photo handling application, I think. In a wider context of applications in general, a selection can, of course, refer to many other things, e.g. an item one selects in a dropdown box.

With this background, I now went to Google and found an image to try out some of the techniques taught, copied it to the clipboard and opened it in AF with "New from Clipboard". This gave me the image layer. I used the paintbrush selection tool to select an area on the canvas and wanted to create a new layer with that selection and therefore used the "Duplicate Selection" function. The whole image layer was copied instead of just the pixels visible inside the selection on the canvas. This did surprise me as it was a different result to all the example in the course that I had tried out, and I thus thought I had done something wrong in using the selection tool, so I repeated the process a few times to no avail. A post in the forum of the course gave no help -- they could not reproduce the effect (probably trying it out on a pixel layer). Eventually, I discovered that the layer in the photo document was of the type "Image" as opposed to other layers that I had been working with. Thus, I right-clicked on the layer and found the "Rasterize" function, which solved the problem. But frustration and waste of time could have been avoided.

Yes, it is obvious that "selection" has a specific context. What is confusing is that when one has an active selection on the canvas, the context suddenly is different depending on whether there is an image layer or a pixel layer as the active one. I do understand that the "running ants" area is on the canvas itself and not in the pixel or image layer. I think it would be more logical that the context of "Duplicate Selection" should be the same for pixel and image layers. Furthermore, I think that many other users with a background of just simple photo apps and perhaps simple pixel layer handling would feel the same.

You disagree, fair enough! 

But do tell me:

If one had an active selection on the canvas based on an image layer (e.g. some area selected with the paintbrush tool) and the "Duplicate Selection" function were to just take the pixels from the image layer (as if it was a pixel layer) and copy these to a new pixel layer, would that spoil or disrupt any specific function or work process in AF?

If not, I think that implementing this change of functionality might save many users time and frustration.

For me, it is no longer an issue as I now understand how AF currently work in this regard.

 

/Frank

 

[R C-R]

22 minutes ago, Frank M. said:

Yes, it is obvious that "selection" has a specific context. What is confusing is that when one has an active selection on the canvas, the context suddenly is different depending on whether there is an image layer or a pixel layer as the active one.

The context is different not just if it is an Image or pixel layer but also if for example it is another layer type like vector or quick shape, or if there is currently no selected layer, or more than one selected layer (which can include several different layer types).

40 minutes ago, Frank M. said:

If one had an active selection on the canvas based on an image layer (e.g. some area selected with the paintbrush tool) and the "Duplicate Selection" function were to just take the pixels from the image layer (as if it was a pixel layer) and copy these to a new pixel layer, would that spoil or disrupt any specific function or work process in AF?

To begin with, a marching ants selection is not in & of itself based on or a part of an image or any other kind of layer. As been mentioned many times now, it just defines an area on the canvas. Note also that the selection tools can make marching ants selections based on the current layer, on no layer, or on multiple layers. Each of the selection tools has options that together with whatever layer type(s) might be selected define the context of what, if anything, will be duplicated. So you cannot assume that there will always be only (or any!) pixels in the area defined by the marching ants selection.

As for 'taking' pixels from an image layer, note that unlike pixel layers, Image layers have their own pixel resolutions independent of the document pixel resolution. Among other things, since more than one Image layer might be selected as well as some other layer type(s), & each of those Image layers might have a different pixel resolution (including different horizontal & vertical resolutions), there is no clearly defined logical way to determine how everything that might be enclosed by a marching ants selection should be combined & duplicated into a single rasterized layer at the document resolution.

There are workflows that would be disrupted if this worked as you suggest, including ones in which the only layer currently selected is an Image type. For example, I sometimes duplicate one (or more) entire image layers at its own 'native' pixel resolution while there is a marching ants selection. I do not want to have to deselect the marching ants selection to do this -- that would add two extra steps to my workflow, one to deselect it & another to reselect it afterwards. I also do not want to rasterize the duplicate(s) because that is a destructive operation.

All this said, I do understand that new users may be unfamiliar with the unique properties of Image layers. I was too when I first started using the Affinity apps. But I have since learned to take advantage of that. I am sure you can too once you understand how more of the various selection types, tools, & other features & options can be combined in different workflows. For that matter, I am still learning new ones based on what other users & staff members have shared in the forums.

It has nothing to do with talking down to anyone. It is just that there are so many different possibilities that there is always something new to discover, & considering a different viewpoint can be a huge help in seeing what some of them might be.

[Move Along People]

-

 

[Frank M.]

On 1/19/2020 at 5:24 AM, haakoo said:

It's Serif's choice to have it implemented this way and their decision if it should change or if there should be a warning or an automate rasterization.

Certainly, it is their decision. 
However, I am also sure that they follow the forum and sometimes base their UI choices on the viewpoints expressed by the users. 

 

On 1/19/2020 at 5:24 AM, haakoo said:

Well you learned, so can others.

, well yes, but could the journey be easier for other new users, that would not be such a bad thing, as long as it does not spoil anything for the power users.
 

/Frank

[Frank M.]

I can't seem to get rid of this duplicate post the forum created when I was answering another post!
 

/Frank

[Frank M.]

On 1/19/2020 at 5:12 AM, R C-R said:

To begin with, a marching ants selection is not in & of itself based on or a part of an image or any other kind of layer. As been mentioned many times now, it just defines an area on the canvas

Yes, I know. However, there is still a quite strong relationship between a canvas selection and layers in the sense that one most often would use a layer as the basis of a selection and not select blindly on a canvas with no layers. In my case, it was the paintbrush selection tool used on an image layer. This worked as expected with areas of the same colour near the paintbrush being selected. Thus, there must be some default rasterisation or conversion taking place for that to work? This is also why it seems so compelling, logical and pleasing that the "Layer" -> "Duplicate Selection" should continue to work the same way, using those same pixels that were the basis of the paintbrush selection process to be copied to a new pixel layer.

 

 

 

 

On 1/19/2020 at 5:12 AM, R C-R said:

There are workflows that would be disrupted if this worked as you suggest, including ones in which the only layer currently selected is an Image type. For example, I sometimes duplicate one (or more) entire image layers at its own 'native' pixel resolution while there is a marching ants selection. I do not want to have to deselect the marching ants selection to do this -- that would add two extra steps to my workflow, one to deselect it & another to reselect it afterwards. I also do not want to rasterize the duplicate(s) because that is a destructive operation.

Right, I see that now.

I experimented with a document with many pixel layers, clicking on several of these (selectin them) while also having an active "marching ants" selection on the canvas. In this case, the "Layer" -> "Duplicate Selection" function copies the layers and does not make a new pixel layer with pixels from the canvas selection. This is also a perfectly logical way for it to work.

So I see now there are two possible and logical functions for this single menu-point.

Perhaps the best would be a new, separate menu-point "Layer" -> "Duplicate Canvas Selection" (inactive when there is no canvas selection). This function would copy the pixels to a new pixel layer. I know you will now say that image layers may have many different resolutions and there may be other layer types :-). However, again, there must be a basic or default pixel conversion or interpretation that combines the layers to show the combined document in AF (when all layers are set visible or the document is exported as a PNG) as well as form the basis of, e.g., the paintbrush selection process. Would it be so wrong for that pixel representation to be used as a basis together with the canvas "marching ants" selection to form a new pixel layer?

In any case, each menu-point would then have a clear, separate function, which is a good thing.
Unless, of course, you now tell me that there are other functions for these that I do not know about as a new user.

Thanks for taking the time to go over these things in AF. It has been very informative :-).

 

/Frank

 

[R C-R]

1 hour ago, Frank M. said:

However, there is still a quite strong relationship between a canvas selection and layers in the sense that one most often would use a layer as the basis of a selection and not select blindly on a canvas with no layers.

There is nothing blind about making a 'marching ants' selection. Even on a blank canvas the marching ants are visible, & the various selection tools can be used to add to, subtract from, or intersect with an existing one. When one of the selection tools is active, the selection can even be resized or moved precisely using the Transform panel.

It is also not necessarily true that everyone would most often make or modify one based on anything in any layer, existing or not. One example is when creating original artwork or a background to place photos on, a 'marching ants' selection might be used to define a specific area of the canvas to fill with brush strokes or stamps like those in the (highly recommended) Frankentoon Concept Master Volume 1 Nature Brush pack. 

2 hours ago, Frank M. said:

This is also why it seems so compelling, logical and pleasing that the "Layer" -> "Duplicate Selection" should continue to work the same way, using those same pixels that were the basis of the paintbrush selection process to be copied to a new pixel layer.

"Layer" -> "Duplicate Selection" does always work the same way. It is just that "selection" does not always refer to the same thing, nor in the greater scheme of things is there any reason that it should.

2 hours ago, Frank M. said:

Perhaps the best would be a new, separate menu-point "Layer" -> "Duplicate Canvas Selection" (inactive when there is no canvas selection). This function would copy the pixels to a new pixel layer. I know you will now say that image layers may have many different resolutions and there may be other layer types :-). However, again, there must be a basic or default pixel conversion or interpretation that combines the layers to show the combined document in AF (when all layers are set visible or the document is exported as a PNG) as well as form the basis of, e.g., the paintbrush selection process.

I think you may be confusing what is rendered to (shown on) the screen with the document's actual content, & likewise with what the various export options support. For reasons both technical & practical, they are not necessarily the same thing, even when working at 'actual' or 'pixel' zoom levels. This also applies to selections made with the Selection Brush Tool & so on -- you cannot safely assume there is one 'basic' conversion/interpretation that always applies to everything, whether just image layers or not.

So while it might be possible to add some menu item that at least more or less works as you envision, it would be much more complicated to implement that you might think, & even so, it is unlikely it would always work as someone else might expect.

[Frank M.]

57 minutes ago, R C-R said:

It is also not necessarily true that everyone would most often make or modify one based on anything in any layer, existing or not. One example is when creating original artwork or a background to place photos on, a 'marching ants' selection might be used to define a specific area of the canvas to fill with brush strokes or stamps like those in the (highly recommended) Frankentoon Concept Master Volume 1 Nature Brush pack

Ok, interesting!

56 minutes ago, R C-R said:

So while it might be possible to add some menu item that at least more or less works as you envision, it would be much more complicated to implement that you might think, & even so, it is unlikely it would always work as someone else might expect

Alright, I understand.

I think this subject has now been fully explored with the many facts covered and viewpoints expressed. 

Thanks, again, for your replies, which have given insights and ideas to work with. I will now go on with AF, gaining more experience. 

 

Best regards,

Frank