Why are pasted pixels not automatically rasterized?

[orangefizz]

I don't know if there's a prefs setting for this but I find myself constantly frustrated that, after pasting an image into AP, I can't select an area of it and copy without it copying the entire image. This added step of first rasterizing the image before you can use parts of it seems unnecessary.

(And if I'm selecting an AREA with the lasso/whatever, doesn't the app understand what my intention is?)

Am I alone in this?

[fde101]

Try Copy Merged (Command+Shift+C on the Mac) rather than just Copy (Command+C on the Mac).

Remember that in the Affinity products there are two selections going on at the same time: the layer selection (which is more vector-oriented and is managed by the Layers panel, move tool, etc.) and the raster selection (managed by the marquee and freehand selection tools, flood select tool, selection brush, etc.).

In spite of its name and focus, Affinity Photo is primarily a vector-based application which happens to have a focus on tools for manipulating the raster content of its layers.  The normal "Copy" command follows that of the other Affinity products and considers the layer selection rather than the raster selection.

Copy Merged pays attention to the raster selection.

[orangefizz]

3 hours ago, fde101 said:

In spite of its name and focus, Affinity Photo is primarily a vector-based application which happens to have a focus on tools for manipulating the raster content of its layers.

This just blew my mind. @Serif: I have so, so many questions.

Thanks for the tip.

[R C-R]

Despite what you might think, in Affinity Image layers are a special kind of container layers that include more than pixels. This includes the image's original color space, plus its resolution and physical dimensions when placed at its default 'native' resolution.

[orangefizz]

44 minutes ago, R C-R said:

Despite what you might think, in Affinity Image layers are a special kind of container layers that include more than pixels. This includes the image's original color space, plus its resolution and physical dimensions when placed at its default 'native' resolution.

I guess I'm just so habituated to Photoshop—since it's earliest incarnations as a pixel mangler—I assumed that a competitor's "Photo" app would be in the same vein. A mental adjustment...

[walt.farrell]

9 hours ago, fde101 said:

The normal "Copy" command follows that of the other Affinity products and considers the layer selection rather than the raster selection.

I don't think that's true. If you have a pixel selection, it will be what's copied. With an active Image layer, though, you don't have a pixel selection, so, lacking a pixel selection, the complete layer is copied.

9 hours ago, orangefizz said:

(And if I'm selecting an AREA with the lasso/whatever, doesn't the app understand what my intention is?)

It has been suggested that there is room for an improvement in the Assistant Manager, such that it could detect actions like this that require access to pixels, and allow you the option of automatic rasterization. And this could be one of them if Serif ever decides to do something like that. But given that we have Copy Flattened (Windows) and Copy Merged (Mac) as fde101 mentioned, which will perform the function without destructively modifying the source layer, I think that might be a low priority.

---

But from your comment, I think there is an aspect of pixel selections in Affinity applications you may not understand. While you may use a Tool to generate a "marching ants" outline, that outline is not the pixels themselves. It is just a separate line drawn on the workspace. It does not even require any data to be in the document. You can use File > New and create a completely empty document, and draw a pixel selection with one of the marquee selection tools, or the freehand selection. There are no pixels there at all, but you have the area marked off.

Later, if you happen to put any pixels in that area, you would be able to copy them. But if you put non-pixel data into the document, and make that non-pixel layer active, there are again no pixels to copy with the standard Edit > Copy.

And if you have multiple pixel layers, and a pixel selection, the pixels it "contains" are from whichever pixel layer happens to be active. The selection is not part of any particular layer. Select one layer in the Layers panel or by clicking on the workspace, you'll get those pixels. Select a different layer, you'll get different pixels. The selection doesn't care what is active, but the Copy operation does.

[orangefizz]

4 minutes ago, walt.farrell said:

But given that we have Copy Flattened (Windows) and Copy Merged (Mac) as fde101 mentioned, which will perform the function without destructively modifying the source layer, I think that might be a low priority.

This is essentially where Smart Objects come in in Photoshop. You'd only rasterize it if you wanted to edit it destructively.

But pasting an image—versus *placing* one—gives you instant access to the pixels.

8 minutes ago, walt.farrell said:

While you may use a Tool to generate a "marching ants" outline, that outline is not the pixels themselves.

As I'm discovering. An app named "Photo" that is actually vector-based... Hm.

[R C-R]

12 minutes ago, walt.farrell said:

Select a different layer, you'll get different pixels. The selection doesn't care what is active, but the Copy operation does.

You can even make a marching ants selection in a new document that as yet has no layers at all! All it does is define an area of the document where there may or may not be any pixels.

5 minutes ago, orangefizz said:

An app named "Photo" that is actually vector-based... Hm.

It isn't just vector based. Like all the Affinity (& many other graphics creation) apps it is a hybrid that can include vectors, text, pixels, images, & so on.

[orangefizz]

45 minutes ago, R C-R said:

It isn't just vector based. Like all the Affinity (& many other graphics creation) apps it is a hybrid that can include vectors, text, pixels, images, & so on.

Yeah I get that, it's just odd that image data is relegated to third class status in a "Photo" app. Guess it's just me...

[R C-R]

2 hours ago, orangefizz said:

Yeah I get that, it's just odd that image data is relegated to third class status in a "Photo" app.

Image data is in no way treated as "third class" in AP. It is just that Image layers include more than pixel image data & as such include some unique features not found in apps that have no equivalent container layer type. 

So when you place an image in AP it is placed as an Image layer. If you then select it with the Move Tool, on the context toolbar you will see several options, among them a "Replace Image" button (which replaces the image at the same size as the one it is replacing) & a drop down menu that allows you to change its resolution proportionally or non-proportionally or return it to its original size. If you rasterize it, you lose the ability to do this & it is converted to the color space of the document.

IOW, it is a non-destructive alternative to automatically rasterizing the image when it is placed in the document & changing its color space.

[orangefizz]

*pixel data

[walt.farrell]

1 hour ago, orangefizz said:

*pixel data

But it's not "pixel data" until you Rasterize the Image layer.

[orangefizz]

1 minute ago, walt.farrell said:

But it's not "pixel data" until you Rasterize the Image layer.

...right, and that you have to first rasterize the data in a "photo" app is exactly my point since post #1! 😃

[walt.farrell]

11 minutes ago, orangefizz said:

...right, and that you have to first rasterize the data in a "photo" app is exactly my point since post #1! 😃

As mentioned before, Photo is not strictly a photo app. It also has vector capabilities, and other non-pixel functions.

This is one of them. It's similar to Placing/Pasting a .afphoto file, or a .afdesign or .afpub file, or a .psd file, into your "photo".

[thomaso]

14 hours ago, orangefizz said:

This is essentially where Smart Objects come in in Photoshop. You'd only rasterize it if you wanted to edit it destructively.

Would you name "Smart Objects" in PS as "third class", too? For me they were always more of an optional enhancement than a declassifying limitation, ... years before Affinity came along.

Maybe you'd like a quick "Rasterize" button?

[orangefizz]

I can right-click and rasterize also, it's just the necessary added step (as with many added steps in AP). When you're dealing with dozens and dozens of images in a session, that tends to slow things down.

Guess I'll post a feature request for a prefs option that automatically rasterizes pasted/dragged images and leaves placed images un-rasterized.

Thanks all.

[thomaso]

I agree that there are an unexpected number of extra steps required in the Affinity user interface, for some I even can't imagine they were done "by design" but resulted from an incomplete concept when coding was started.

For your feature request you might also consider the Assistant options that already offer some "always Rasterize automatically" settings, I wonder why there is no auto-rasterize option for a current pixel selection + e.g. cut or copy with an image / vector layer selected.

[R C-R]

38 minutes ago, walt.farrell said:

But it's not "pixel data" until you Rasterize the Image layer.

Image layers do include pixel data. It is just that they also contain other data that Pixel layers do not include, like color space & 'native' resolution data. So, two types of layers, each with its own functionality & uses.

[R C-R]

1 hour ago, orangefizz said:

I can right-click and rasterize also, it's just the necessary added step (as with many added steps in AP). When you're dealing with dozens and dozens of images in a session, that tends to slow things down.

Just curious but how often do you want to copy just parts of more than a few of your dozens of pasted images?

[thomaso]

1 hour ago, orangefizz said:

When you're dealing with dozens and dozens of images in a session, that tends to slow things down.

I can't imagine your workflow / document but I assume that "dozens and dozens of images" in 1 document cover each other and thus require a lot more additional clicks than just to get rasterized.

In case you need only partial pixel selections of these images: If you open all of them then ever image appears as separate tab, can get accessed individually via the View or Windows menu list of open files + gets opened as pixel layers which allows you to select and copy directly and paste these parts into your main compositing document.

In case you prefer dragging to get images into Affinity: If you drag images not on a canvas but on the UI then they get opened, not placed and thus are pixel not image layers.

[orangefizz]

52 minutes ago, R C-R said:

Just curious but how often do you want to copy just parts of more than a few of your dozens of pasted images?

often enough that I'm posting here brother

[orangefizz]

17 minutes ago, thomaso said:

I can't imagine your workflow / document but I assume that "dozens and dozens of images" in 1 document cover each other and thus require a lot more additional clicks than just to get rasterized.

Not all in one document necessarily but repeatedly when comping—pasting images from other sources.

18 minutes ago, thomaso said:

In case you need only partial pixel selections of these images: If you open all of them then ever image appears as separate tab, can get accessed individually via the View or Windows menu list of open files + gets opened as pixel layers which allows you to select and copy directly and paste these parts into your main compositing document.

Weird! When I open images and use the marquee to select an area to copy or Cmd-J, it works as expected...it sees pixels. But if I drag an image onto the canvas it's seen as a different object, same when I paste an image from the clipboard. (I also swear the new beta update is working differently than it was yesterday but maybe that's just my brain)

22 minutes ago, thomaso said:

In case you prefer dragging to get images into Affinity: If you drag images not on a canvas but on the UI then they get opened, not placed and thus are pixel not image layers.

Also surprising; evidently the decision was made to treat one case as pixels and the other as an image object. My issue here I guess is that pasting some random jpeg from the clipboard should not be considered placing. If I want something placed—for non-destructive editing—I'll use the place option.

Good info tx

[orangefizz]

1 hour ago, R C-R said:

Image layers do include pixel data. It is just that they also contain other data that Pixel layers do not include, like color space & 'native' resolution data. So, two types of layers, each with its own functionality & uses.

"pixel data" is not pixels. My whole point in posting here is in trying to decipher when AP gives me direct, unmolested access to the !@(^#&*# pixels in an image.

[R C-R]

3 minutes ago, orangefizz said:

"pixel data" is not pixels.

Then what is it if not data about the pixels in the image????

[thomaso]

22 minutes ago, orangefizz said:

My issue here I guess is that pasting some random jpeg from the clipboard should not be considered placing.

People who use APhoto for page layouts (postcards, flyers, posters etc.) may appreciate this placing behaviour and result. The various text + vector features seem to have influenced this decision and apparently the goal was not an app purely focused on image editing.

Maybe your feature request asks for a modifier key to cause directly a pixel placement? – Did you note that there is one when dragging more than 1 image at a time: pressing option/alt opens a "Place Images" Panel that gives you a choice of the placing order and size for each image of the dragged bunch. – So, it could be easy to code another modifier key that causes rasterisation after placing. – For some users such an easy-access solution might be more useful than a general application or document preference for this feature.