Christoph Werner

The showcased snarkiness is troublesome. It doesn´t help that you apparently have no idea nor any fixes or good workarounds.
The recommendation of using Photoshop is also of little value in this 11 pages spawning thread.
16 bit has 65536 colours.

IF you know a way, please make a Video tutorial explaining the process on how it works.
Bonus points if you show the exporting from Affinity Photo.

Edit;
I will make it even easier for you, here are provided test files to use for each channel.
A.tgaB.tgaG.tgaR.tga
 

I know this way. But the most important for many people: You can't paste copied image information into the active mask layer. I would like to know the reason for this missing function?

I'm for.
Meanwhile, you can try this method, I have been using it since version 1.8


Preparing the TGA file for editing
Move the alpha channel to the new mask layer.
1. Open the file for editing and make sure the document background is transparent.
2. Select the background layer in the Layers panel.
3. Create a mask from Background Alpha in the "Channels" pane.
Removal of the original alpha and discovery of RGB pixels:
4. Select the Background layer again;
5. Select the fill option for "Background Alpha" in the "Channels" pane.
6. Now you can hide the mask layer and check if information in RGB channels is preserved.
Editing the mask.
7. After the mask layer is displayed again, you can start editing.
ATTENTION!!!!!!!
Keep the mask position on the top layer and under no circumstances merge it with other layers or flatten the file before exporting. In Affinity Photo, it seems that masks remove information about other channels from the layer they are pinned to, so be sure to keep the mask on a separate layer.
Export to TGA
8. Make sure the background is transparent in the export options.
When re-editing the file, follow the instructions provided.
Unfortunately, working in this way is a bit tedious but not impossible.
You can speed up the preparcess by using macro.
Textures.afmacros

No argument here, just saddens me that this feature is likely the most requested at this point on the forum and the devs have simply ignored us for years at this point. I absolutely love AP otherwise.

I wasn't meaning to imply those features were unimportant - they are important for me too. I'm mainly working in 3D CGI - so 32bit renders, ACES and EXR are one of the main reasons I got AP in the first place.  It just seems very odd to me that they claim their target market to be 'photographers and photo editing', but do support advanced features and functionality such as 32b, ACES, EXR - and yet don't support basic features like direct and unrestricted alpha manipulation. Bizzare.

I'm working with 32 bit EXR files, even multilayered... So it's not that unimportant. But actually this features are needed for compositing tasks mainly, where AP doesn't make much sence.

The development seem to be more interested in "new cool" features instead of boring industry standards...
But I'm still hopeful while I'm back using photoshop meantime.

The development seem to be more interested in "new cool" features instead of boring industry standards...
But I'm still hopeful while I'm back using photoshop meantime.

The development seem to be more interested in "new cool" features instead of boring industry standards...
But I'm still hopeful while I'm back using photoshop meantime.

Alpha channel editing is something the most professionals do every day. And the confusing/unusable system in Affinity is something that should be chagend asap in my opinion. I also don't understand why this is not touched for such a long time. AP could be so good.
I stopped to use AP because of this and jumped back to Photoshop.

Alpha channel editing is something the most professionals do every day. And the confusing/unusable system in Affinity is something that should be chagend asap in my opinion. I also don't understand why this is not touched for such a long time. AP could be so good.
I stopped to use AP because of this and jumped back to Photoshop.

Well, this topic is more over 3 years old now. I'm not surprised that people are still trying to find answers.
As far as I know, there never was any response nor reaction from official site and probably never will be.
I guess the very most of us gave up on this "already".

I'm just still following this topic, because... well, hope dies last... even though it is almost burned out

But the truth seems to be:
They don't answer, they don't care = Do not await this feature to be present at any time in any future

I know this, as I wrote it above. But in my opinion it must be always ensured the user can continue the work in some way. Even if the file size will be increased a bit, the current option to save extremely low quality thumbnails can still exists, but there should be an option to save "rasterized previews" of linked layers, too.
I've prepared an example to document my proposal
In the image below you see a linked photo file. On the left you see what happen, if the file would be rasterized, but still missing. This is my proposal. On the right side you see, what Affinity Photo is doing today if a linked file is missing.
Here a list to compare the Affinity Photo project file sizes on a harddrive:
File size using linked image (Current Affinity way) = 0.739 Megabytes File size using rasterized image preview (My proposal) = 1.7 Megabytes File size with embedded images (Affinitys way, if we embed the photo. What I don't want to and isn't my proposal!) = 5.77 Megabytes So you can see the project files size increases if we save rasterized linked layer previews, but it's still smaller than embedding it in the default way. My way is safer, because it always keeps a backdoor to work with the project, even if there are file links missing.

And following an example how my solution could be implemented into Affinity Photo:
The exclamation mark warns the user of missing data. You can see it in the layer panel and at the related image position. If the user ignores the warning and want to change the layer, then always the existing layer resolution will be used for further changes. A further proposal: As long as the user doesn't save the project, it will keep the linking information of missed files. So if the user relinks the missing data later, any afterward changes will be executed for the relinked files. That would improve the work with Affinity Photo a lot in my opinion.
The developers could implement my idea as an additional option. There is no need to have just one way.

What?! Sounds crazy, but works.
Never thought I can do this with the gradient tool (I come from Photoshop).
Thank you a lot!

Completely agree. Editing alpha channels/masks in Affinity Photo is still a no go, if you come from Photoshop.
Sorry for this, but why the hell I can't just edit an alpha channel like a regular pixel channel?! And I don't mean to paint something with a brush in the alpha/mask.
I mean functions like changing color levels, curves or adding blur... Things that works like a charm in Adobe Photoshop.
Don't misunderstand me: I like Affinity Photo a lot and try to work in this tools as much as possible. But alpha editing is still one thing that really su...s. Or I didn't understand until today, how Affnitys solution is better. What solution it ever is, to change levels in alpha channels in some seconds.
😪

You're right... it works! THANK YOU!

I meant the drop down in the Dynamics tab, 

That isn't the point, which is that the user has to remember how much it was scaled from its original size for that to work because the apps do not.

But are they not destructive in the sense that there is no way to return them to the original size once rescaled?

There is no Corner Tool in Photo.
One, quite horrible, method in Photo is to create two new nodes near the corner, then alter the corner node to Smooth, then move the corner node inwards – with probably some extra manipulation of the control handles.
See my attached video where I do it very quickly. You will probably need to take much more time over it to make something look good.
Hopefully someone else can show you a better/nicer method.
P.S. Photo isn’t really the best application for vector manipulation as it’s more suited for image manipulation.
2021-07-12 08-58-49.mp4

Welcome to the Affinity forums 
Something that works currently (AP 1.9.3), is to ensure an object that represents alpha is at the top of the stack when you export to an image file.
That object can be a Pixel Mask or it can be anything with its intensity inversely mapped to opacity (black through white mapped to opaque through transparent) and blend mode set to Erase. The mapping can be done with a downward ramp in the left-hand graph of the object's Blend Options.
Below are 3 documents which all export to PNG without RGB being destroyed where alpha is zero.
1.  alpha by erasing.afphoto

 
2. alpha by erasing (simplified).afphoto

 
3. alpha by masking.afphoto

 
Screenshot of a PNG that was exported from from "alpha by erasing.afphoto": 

 
Screenshot of that PNG after filling alpha to prove the RGB was not destroyed:

 
 
 

This is not quite the same thing as your request, OP, but it is related to Embedded Documents.
One aspect I want to see changed with Embedded Documents is the auto-updating of the main document that is happening while working inside one of those documents. I've compared Embedded Documents with Smart Objects in Photoshop and the performance difference is pretty significant. In Photoshop I can save my progress while inside a Smart Object, but in Photo I have to switch back to the main document to save. Because of the auto-update feature the switch can be heavily delayed in big projects or even freeze, so you end up losing all the changes done to the embedded document. With Photoshop Smart Objects this does not happen. Updates to the main document only happen when you save your Smart Object and switching between the two documents takes very little time, thus making it a very good performance booster for larger projects where the amount of layers can blow up quite a bit.
Also, I have noticed that Embedded Documents are really resource heavy even while working on the main document, while in Photoshop you in fact save performance by converting stacks of layers into a Smart Object layer. Photo needs a way to create an option for Embedded Documents which is resource light and does not automatically update your main canvas until the user decides that it is time to save.

Perhaps because I work alone since 1988, and that's for a reason…
It's OK, sure. An nice option to have when needed. But I, for one, don't see it as "essential", and probably wouldn't ever use it at all.

Then the user should ultimately know what they are doing:
either keep track of your linked documents and their location or if you aren't capable of that for whichever reasons, embed your linked document and stop worrying That said…
… I think it's a good idea.
Having options is usually a Good Thing™.