Affinity Photo 2.1 (not sure if it is only related to this) copy and paste and export is not consistent

[R C-R]

4 minutes ago, lepr said:

 For 8 pages now you have exhibited a mental block preventing you from being able to grasp the concept of the Pixel Selection itself being a raster object, or in other words, a rows by columns grid of pixels. 

Nonsense! The marching ants selection is not actually a raster or any other kind of object. It is just a defined area of the document's pixel grid, one that may or may not enclose any full or partial pixels of any layer the document might include. it can be manipulated independently of any actual object the document might contain & should not have any effect on any layer unless/until it is applied to that layer.

14 minutes ago, lepr said:

Every time I mention the raster grid or pixels of the Pixel Selection, you mistakenly interpret that as the raster grid or pixels of a Pixel Layer that is also in the document.

Nope. I have been as clear as I possibly can be that what you call "the raster grid or pixels of the Pixel Selection" is just the marching ants pixel selection, & it is totally independent of any pixel layer that may or may not be in the document. Why you never seem to understand that I am as careful as I can be to make this distinction is a mystery to me.

[R C-R]

Why?.afphoto is an example of what puzzles me about the creation of partially transparent pixels during the copy & paste when everything is perfectly pixel aligned & the only pixel layer's content is 100% opaque.

History is included so you can follow step by step what I did to insure to the extent possible that there was no partial transparency to begin with & that the marching ants selection what perfectly pixel aligned both before & after resizing it. (There is a copy step but it does not seem to be recorded as a separate step.)

Again, it seems to me that it should be trivially easy to copy something like I have done in that example without its pixels being altered in any way

[lepr]

3 hours ago, R C-R said:

Nonsense! The marching ants selection is not actually a raster or any other kind of object [... blah blah ...]

 

Watch the attached video where I start with an empty document then:

1. add an Empty Mask to the document
2. use a soft white brush to paint various values into the pixels of the Mask
3. option-click the Mask thumbnail to visualise the Mask as a monochrome image
4. option-click the thumbnail again to remove the visualisation
5. cmd-click the thumbnail to perform Selection From Layer command - the created Pixel Selection (PS) is visualised by marching ants drawn in the view
6. right-click the thumbnail and pick Delete to destroy it - there is now nothing in the Layers panel to pollute your judgement of the following
7. click the Quick Mask button with 'Show Mask As Black' mode to visualise the PS as a monochrome image - it has various values and is identical to the  monochrome visualisation of the Mask from which the Pixel Selection was made
8. use the handles of the PS's bounding box to stretch, scale and shear it - non-destructive transformations of the PS's raster grid are clearly taking place (in other words, no pixel values are changing) and its pixels are clearly independent of the document raster grid
9. toggle Quick Mask off to show the marching ants enclose only the PS pixels which have a value of greater than 50% and see how some of the PS exists outside the marching ants
10. toggle Quick Mask on and reset the PS's rotation and shear to zero, restore its width and height to original measures, and snap it to the document raster grid
11. exit Quick Mask mode to see the marching ants visualisation again
12. click the 'add Mask' button to create a new Mask from the PS
13. option-click the Mask thumbnail to visualise the Mask as a monochrome image and see that it matches the original Mask

If you now accept the Pixel Selection is a raster object with transformation matrix, maybe we can move on with educating you about the reason for the annoying feathering that occurs when you copy a Pixel Layer while a stretched Pixel Selection is active.

If you are still in denial, you'll be unable to understand the reason for the feathering.

 

 

[R C-R]

@lepr, drawing on a mask with a soft brush to create partial mask transparency has nothing to do with what I was asking about when no pixels in the layer the marching ants selection is applied to include any transparency, and the marching ants selection only encloses those pixels when the copy is made.

To repeat, nobody is saying that what you demonstrated does not occur if not all the selected pixels have a value of greater than 50%. We all agree on how that works. But the issue is what happens when all the enclosed pixels of the layer are fully opaque, & why the copy includes ones that are not.

[CyberAngel]

5 hours ago, lepr said:

The Pixel Selection's transformation matrix is only ever applied to the Pixel Selection's own raster grid of pixels. For 8 pages now you have exhibited a mental block preventing you from being able to grasp the concept of the Pixel Selection itself being a raster object, or in other words, a rows by columns grid of pixels. Every time I mention the raster grid or pixels of the Pixel Selection, you mistakenly interpret that as the raster grid or pixels of a Pixel Layer that is also in the document. As long as you have that mental block, you will think I am talking nonsense.

As for the rest of your message from which I quoted, there is no point in me explaining anything to you further until you deal with that mental block.

Right, and I have said that the expectation is that we should be able to resize the Marquee without it doing what it is doing. I frankly do not care what it is doing now, only what the expectation of how everything else works. In that I mean I can use the transform to resize everything without affecting what is underneath, yet the marquee seems to be different.

That is all that matters here.

[CyberAngel]

2 hours ago, lepr said:

 

Watch the attached video where I start with an empty document then:

1. add an Empty Mask to the document
2. use a soft white brush to paint various values into the pixels of the Mask
3. option-click the Mask thumbnail to visualise the Mask as a monochrome image
4. option-click the thumbnail again to remove the visualisation
5. cmd-click the thumbnail to perform Selection From Layer command - the created Pixel Selection (PS) is visualised by marching ants drawn in the view
6. right-click the thumbnail and pick Delete to destroy it - there is now nothing in the Layers panel to pollute your judgement of the following
7. click the Quick Mask button with 'Show Mask As Black' mode to visualise the PS as a monochrome image - it has various values and is identical to the  monochrome visualisation of the Mask from which the Pixel Selection was made
8. use the handles of the PS's bounding box to stretch, scale and shear it - non-destructive transformations of the PS's raster grid are clearly taking place (in other words, no pixel values are changing) and its pixels are clearly independent of the document raster grid
9. toggle Quick Mask off to show the marching ants enclose only the PS pixels which have a value of greater than 50% and see how some of the PS exists outside the marching ants
10. toggle Quick Mask on and reset the PS's rotation and shear to zero, restore its width and height to original measures, and snap it to the document raster grid
11. exit Quick Mask mode to see the marching ants visualisation again
12. click the 'add Mask' button to create a new Mask from the PS
13. option-click the Mask thumbnail to visualise the Mask as a monochrome image and see that it matches the original Mask

If you now accept the Pixel Selection is a raster object with transformation matrix, maybe we can move on with educating you about the reason for the annoying feathering that occurs when you copy a Pixel Layer while a stretched Pixel Selection is active.

If you are still in denial, you'll be unable to understand the reason for the feathering.

 

 

 

That video does nothing for almost half of it, but what does it have to do with the marquee tool!

[R C-R]

4 minutes ago, CyberAngel said:

Right, and I have said that the expectation is that we should be able to resize the Marquee without it doing what it is doing.

I would be curious to know if @lepr has any thoughts on how that could be done or why it can't be done now, particularly when working with something like my Why?.afphoto file where to me there still seems to be no reason for any pixels to be made partially transparent when copied, or in your original test file.

[CyberAngel]

2 minutes ago, R C-R said:

I would be curious to know if @lepr has any thoughts on how that could be done or why it can't be done now, particularly when working with something like my Why?.afphoto file where to me there still seems to be no reason for any pixels to be made partially transparent when copied, or in your original test file.

No idea, I am still learning the suite of tools.

I can only say what I have expected to happen, coming from the usage of other programs. And why I believe this is an issue. While I accept there are work arounds, they are all in my opinion long winded unnecessary steps to something that appears to be broken.

[R C-R]

2 minutes ago, CyberAngel said:

While I accept there are work arounds, they are all in my opinion long winded unnecessary steps to something that appears to be broken.

FWIW, it appears to be broken to me, too. I can think of no good reason why, when everything is pixel aligned, a copy should not create an exact copy of what the marquee applied to a pixel layer includes.

[CyberAngel]

6 minutes ago, R C-R said:

FWIW, it appears to be broken to me, too. I can think of no good reason why, when everything is pixel aligned, a copy should not create an exact copy of what the marquee applied to a pixel layer includes.

Exactly.

[lepr]

7 hours ago, R C-R said:

@lepr, drawing on a mask with a soft brush to create partial mask transparency has nothing to do with what I was asking about when no pixels in the layer the marching ants selection is applied to include any transparency, and the marching ants selection only encloses those pixels when the copy is made.

LOL! You've completely misunderstood the point of that video, too. Incredible!

It was precisely to counter your exclamation of "Nonsense! The marching ants selection is not actually a raster or any other kind of object." I quoted that specific statement at the start of the message.

It shows the Pixel Selection being non-destructively transformed because it has a transformation matrix.

It shows the Pixel Selection is a raster object - a grid of pixels - and the marching ants are one type of visualisation of it, and a monochrome image is another type of visualisation of it.

If I had created a Pixel Selection with the Rectangular Marquee Tool, the object being transformed in Quick Mask could have been mistaken for for a vector rectangle, so I deliberately created a Pixel selection with a distinctly pixelated appearance in Quick Mask.

 

7 hours ago, R C-R said:

To repeat, nobody is saying that what you demonstrated does not occur if not all the selected pixels have a value of greater than 50%. We all agree on how that works. 

We agree the marching ants show a boundary between selection values of <=50% and >50%. That is all we agree on.

If as you claim, "the marching ants selection is not actually a raster or any other kind of object", where do you suggest these values ranging from 0% through 100% are stored?

 

7 hours ago, R C-R said:

But the issue is what happens when all the enclosed pixels of the layer are fully opaque, & why the copy includes ones that are not.

As I said, it won't be possible to understand the explanation when you cannot accept the concept of the Pixel Selection being an object. I briefly explained it to Brian earlier, but that would have been incomprehensible to you because of your current mindset.

 

[lepr]

5 hours ago, CyberAngel said:

Right, and I have said that the expectation is that we should be able to resize the Marquee without it doing what it is doing. I frankly do not care what it is doing now, only what the expectation of how everything else works. In that I mean I can use the transform to resize everything without affecting what is underneath, yet the marquee seems to be different.

That is all that matters here.

So, you are not interested in how the Pixel Selection functionality is currently implemented and why the feathering problem is happening. You just want the software changed  to eliminate the feathering problem. That's fine. There's nothing wrong with that attitude. I have been posting for people who do have an interest in why things are as they are.

[lepr]

5 hours ago, R C-R said:

I would be curious to know if @lepr has any thoughts on how that could be done or why it can't be done now, particularly when working with something like my Why?.afphoto file where to me there still seems to be no reason for any pixels to be made partially transparent when copied, or in your original test file.

Yes, I do have ideas. However they will seem like nonsense to you, as you've stated several times. If what I write about the current situation is nonsense, my suggestions would similarly look nonsensical, surely.

[R C-R]

11 hours ago, lepr said:

It shows the Pixel Selection is a raster object - a grid of pixels - and the marching ants are one type of visualisation of it, and a monochrome image is another type of visualisation of it.

Whether you consider it an independent raster object in its own right or not, that has nothing to do with nor does it explain why when all the enclosed pixels of the layer an unrotated rectangular marquee selection is applied to are 100% opaque & the selection is precisely aligned to pixel boundaries even after it has been transformed, copying that selection results in a copy in which some of the pixels are less that 100% opaque.

In fact, if you use the Quick Mask method to transform the selection's size or position after first creating it, you can easily make sure it is still precisely aligned to whole pixel boundaries by using the Transform panel. Yet, if you then use that selection to copy some part of a pixel layer that has nothing but 100% opaque pixels, you still get some pixels that are less that 100% opaque.

The implication here is that there is a bug in the implementation of this kind of transformation that fails to include all of each of the pixels along the edges of the selection, as one would expect. It has been suggested that this is, or is at least related to, a bug in what the Transform panel shows in this situation; that even when it displays whole pixel values & the apps are set to show fractional pixels, there are (perhaps due to rounding errors?) still small fractions being selected.

I do not know if that is the source of this behavior or not; only that it makes it much, much harder than it should be to rely on resizing or repositioning a marching ants selection to make an accurate copy of the pixels of some pixel layer it encloses.

[R C-R]

12 hours ago, lepr said:

So, you are not interested in how the Pixel Selection functionality is currently implemented and why the feathering problem is happening. You just want the software changed  to eliminate the feathering problem.

So can we at least agree that this is a problem?

[lepr]

14 hours ago, R C-R said:

Whether you consider it an independent raster object in its own right or not, that has nothing to do with nor does it explain why when all the enclosed pixels of the layer an unrotated rectangular marquee selection is applied to are 100% opaque & the selection is precisely aligned to pixel boundaries even after it has been transformed, copying that selection results in a copy in which some of the pixels are less that 100% opaque.

In fact, if you use the Quick Mask method to transform the selection's size or position after first creating it, you can easily make sure it is still precisely aligned to whole pixel boundaries by using the Transform panel. Yet, if you then use that selection to copy some part of a pixel layer that has nothing but 100% opaque pixels, you still get some pixels that are less that 100% opaque.

The implication here is that there is a bug in the implementation of this kind of transformation that fails to include all of each of the pixels along the edges of the selection, as one would expect. It has been suggested that this is, or is at least related to, a bug in what the Transform panel shows in this situation; that even when it displays whole pixel values & the apps are set to show fractional pixels, there are (perhaps due to rounding errors?) still small fractions being selected.

I do not know if that is the source of this behavior or not; only that it makes it much, much harder than it should be to rely on resizing or repositioning a marching ants selection to make an accurate copy of the pixels of some pixel layer it encloses.

 

Ah, so you won't outright admit it, but you do now accept what you've argued against (and shouted "Nonsense" at for pages: the Pixel Selection is a raster object.

The feathering when copying is performed when the Pixel Selection is active... I will explain this evening since you might be able to understand it now.

 

[CyberAngel]

5 minutes ago, lepr said:

 

Ah, so you won't outright admit it, but you do now accept what you've argued against (and shouted "Nonsense" at for pages: the Pixel Selection is a raster object.

Yet, we can agree that is how it is working. What we disagree on here, is how it should work. You say it is a pixel selection, I say it should not be anything till you take action on it!

 

[R C-R]

5 hours ago, lepr said:

Ah, so you won't outright admit it, but you do now accept what you've argued against (and shouted "Nonsense" at for pages: the Pixel Selection is a raster object.

No. I still do not consider marching ants selections to be independent raster objects like the pixel layer objects they can be applied to. I'm just tired of arguing about it with you.

Nevertheless, I look forward to your explanation of why feathering occurs only when the marching ants selections are transformed after being created, even when everything is pixel aligned & the transformation occurs when no pixel layer is selected.

[Old Bruce]

1 hour ago, R C-R said:

Nevertheless, I look forward to your explanation of why feathering occurs only when the marching ants selections are transformed after being created, even when everything is pixel aligned & the transformation occurs when no pixel layer is selected.

I have been thinking about this and have no idea why the unexpected and unwanted feathering/altering of what is basically an area occurs. It would be interesting to actually find out what the mechanics of it are. But if I never find out I won't lose any sleep over it.

[lepr]

8 hours ago, CyberAngel said:

Yet, we can agree that is how it is working. What we disagree on here, is how it should work.

Every time you respond to my comments, it is blatantly clear that you have completely misunderstood me. Agreement/disagreement is meaningless in this situation.

 

8 hours ago, CyberAngel said:

You say it is a pixel selection, I say it should not be anything till you take action on it!

That sentence looks like gibberish to me.

[R C-R]

1 hour ago, Old Bruce said:

I have been thinking about this and have no idea why the unexpected and unwanted feathering/altering of what is basically an area occurs.

That may help explain why I disagree with @lepr about a pixel selection being an independent raster object. Image a 10 x 10 px document. If I select for instance 10 of its 100 pixels, I do not believe that action, in & of itself, makes those 10 pixels a new raster object. The selection just defines an area of (or marks off a boundary in, if you prefer) the document, nothing more.

[Old Bruce]

2 minutes ago, R C-R said:

Image a 10 x 10 px document. If I select for instance 10 of its 100 pixels, I do not believe that action, in & of itself, makes those 10 pixels a new raster object. The selection just defines an area of (or marks off a boundary in, if you prefer) the document, nothing more.

I agree 100% with this definition of a Marquee selection. The curious part is why there is some sort of distortion happening if we stretch* the area. 

 

* My limited tests showed me no "obvious" distortion happening when I shrunk a marquee selection.

[R C-R]

4 minutes ago, lepr said:

Every time you respond to my comments, it is blatantly clear that you have completely misunderstood me. Agreement/disagreement is meaningless in this situation.

The disagreement is about it should work, not how it does work, & several others besides myself & @CyberAngel believe there is no good reason why it works like it does now.

[lepr]

3 minutes ago, R C-R said:

That may help explain why I disagree with @lepr about a pixel selection being an independent raster object. Image a 10 x 10 px document. If I select for instance 10 of its 100 pixels, I do not believe that action, in & of itself, makes those 10 pixels a new raster object. The selection just defines an area of (or marks off a boundary in, if you prefer) the document, nothing more.

We already know that you utterly failed to comprehend the video I provided you along with a written list of the steps I took. 

[lepr]

6 minutes ago, R C-R said:

The disagreement is about it should work, not how it does work

Whaaaaat!?!? You've written message after message in disagreement with how I said it does work. I have said nothing about how I think it should work. This thread is absurd. I'm trying to communicate with a jackass.