Is there a better workflow I should be using than "rasterize" layer? Making new layer from a complex selection.

[Bedwyr]

I was wondering if there was a different workflow as a best practice than rasterizing an image layer before copying a selection and making a new layer out of it. Or is rasterizing always a good practice when I want to, say, cut out words or elements of images and work with that selection specifically?

References to videos or articles about common operations like this are welcome. This is just what I sort of hit upon for myself in the interim. If there's better ways to do things I can be taught.

Edited June 21, 2024 by Bedwyr
Clarification.

[GarryP]

In order for the pixel selection tools to work with a layer, the layer must be a Pixel layer.

Rasterising is usually the easiest/quickest way to make a layer into a Pixel layer.

There may be better ways to do what you want but we would need specific details to be able to tell if there was a better way for any particular case.

[Old Bruce]

15 hours ago, Bedwyr said:

I was wondering if there was a different workflow as a best practice than rasterizing an image layer before copying a selection and making a new layer out of it.

For this you could try using Edit > Copy Merged.

[Ldina]

14 minutes ago, Old Bruce said:

For this you could try using Edit > Copy Merged.

I agree. You can also do a Layer > Merge Visible (Cmd-Opt-Shift-E on Mac), which will create a NEW pixel layer of all visible layers and leave your original layers intact. Then you can use your pixel selection tools on the merged pixel layer to make selections, copy them, etc. I often do that, and when I no longer need the 'merged pixel layer', I delete it to keep file size reasonable. 

[Bedwyr]

Thanks. Let me ask a slightly deeper question. Are all selection tools in Affinity pixel selection tools?

[Ldina]

@Bedwyr The selection tools allow you to to create a selection, even from an image layer. For example, I usually develop my RAW images in the Develop Persona, and bring them into Photo as "Linked" layers to reduce file size. That layer is considered an "image" layer and is shown as such in the Resource Manager (Window > Resource Manager). I can use the various selection tools (Freehand selection, rectangular marque, Selection brush, flood selection, etc) to create a selection, then I can utilize that selection to create a mask for a Curves, Levels, HSL or other adjustment layer. This will limit the adjustment layer or filter to that portion you have selected. 

BUT...if I try to copy that "selection" from an Image Layer to another layer, it will copy the entire layer rather than just the selection, since the source layer is an "image" layer and is considered immutable. Pixel layers can be modified, but image layers cannot. If I want to copy JUST the pixels in my selection to its own layer, I need to copy it from a "Pixel layer". So, it appears that if you want to copy 'some' pixels from one layer to another, you have to make sure the source is a pixel layer.

Others here have more experience than I do, and may be able to provide a better and more complete explanation, but that is my experience. 

[walt.farrell]

2 minutes ago, Ldina said:

So, it appears that if you want to copy 'some' pixels from one layer to another, you have to make sure the source is a pixel layer.

Or use one of the other methods described above.

[Bedwyr]

16 minutes ago, Ldina said:

@Bedwyr The selection tools allow you to to create a selection, even from an image layer. For example, I usually develop my RAW images in the Develop Persona, and bring them into Photo as "Linked" layers to reduce file size. That layer is considered an "image" layer and is shown as such in the Resource Manager (Window > Resource Manager). I can use the various selection tools (Freehand selection, rectangular marque, Selection brush, flood selection, etc) to create a selection, then I can utilize that selection to create a mask for a Curves, Levels, HSL or other adjustment layer. This will limit the adjustment layer or filter to that portion you have selected. 

BUT...if I try to copy that "selection" from an Image Layer to another layer, it will copy the entire layer rather than just the selection, since the source layer is an "image" layer and is considered immutable. Pixel layers can be modified, but image layers cannot. If I want to copy JUST the pixels in my selection to its own layer, I need to copy it from a "Pixel layer". So, it appears that if you want to copy 'some' pixels from one layer to another, you have to make sure the source is a pixel layer.

Others here have more experience than I do, and may be able to provide a better and more complete explanation, but that is my experience. 

Thank you; that actually helps a great deal. Partly this was a line of questioning so I can better understand the philosophy of the UI. A lot of people just want to get to "how do I make thing happen", but that only goes so far. Knowing *why* it works that way helps the mental model immeasurably. Different needs at different times.

Edit, further question: I assume if I want to preserve the image layer as is, I should just duplicate, make the second layer a pixel layer and edit that one.

[nickbatz]

Usually I use Merge Visible, make the selection and invert it, and then delete what I don't want.

But sometimes I create a new pixel layer and use the Clone Brush tool on it (with Current Layer and Below selected in the context menu) to copy what I want. If you don't need the copied selection to be in the exact same location, it can be easier than dealing with a Merge Visible layer.

[Ldina]

@Bedwyr OldBruce's method works great. You make your selection on the "image layer" or pixel layer, then Edit > Copy Merged. Then "paste" (Edit > Paste, or Cmd-V) and it will paste the selection on a new pixel layer. 

Nicebatz's method works great too. Create a new blank Pixel layer, and use the clone brush with that layer selected, with the context menu set to 'Current Layer and Below'. That will add whatever you clone to the new pixel layer. 

Both methods, and the one I posted previously, will leave your original layer(s) intact and unmodified. I like both of these methods and learned something new myself. Thanks!

 

[Ldina]

@Bedwyr 

21 hours ago, Bedwyr said:

Partly this was a line of questioning so I can better understand the philosophy of the UI.

Here's some additional information you might find helpful. 

In the Selection menu, you have the ability to select Color Range, Tonal Range, Sampled Colors and Alpha Range (see screenshot). In order to use the first three, you need to sample from a "Pixel Layer". It doesn't work on Image Layers, (which are considered immutable unless rasterized first) because these selection options need to be able to see the values of each pixel in order to select something.

Alpha Range will work on Pixel Layers (if you have both opaque and transparent pixels in the layer), and selections by Alpha Range will also work on Filter layers, Adjustment Layers and Masks, since they are really pure alpha (at least that's my understanding...someone please correct me if I am mistaken). 

Attached are two videos by James Ritson (Affinity Photo product expert) which address selecting by Sampled Color and  Pixel vs Image Layers, which you may find helpful. These videos were from Version 1 of the software, so there are a few minor differences, especially in the look of the Layer Panel. Also, v2 introduced the ability to bring linked or embedded "Image layers" into Photo from the Develop Persona, whereas I believe v1 always delivered Pixel Layers into Photo. FWIW, I almost always "link" RAW images during development, which reduces file size in Photo and allows me to go back and forth to readjust my RAW Development settings at any time. Embedded Files also allow back and forth trips to the Develop Persona. Linked files work great, but if you need to share your AP file with others, you may prefer to deliver a pixel layer to Photo, or embed the RAW file. 

Hope this helps further you understanding. I also learned a thing or two in this thread.