NotMyFault

When talking about selections it is important to explicitly state the context: pixel selections: limit 2D area for drawing tools like brushes Active layer selection in layer stack: for layer operations like moving in layer stack, and choose which layers get impacted by brushes or destructive filters (multiple at the same time possible) My statement above are only high level and not fully accurate. The details are more complex, there are tutorials, help files etc explaining all the details. https://affinity.help/photo/English.lproj/pages/Selections/selections_create.html https://affinity.help/photo/English.lproj/pages/LayerOperations/select.html

Pixel Selections are agnostic to layers (independent) once created. it only depends on the selected tool and it’s options (e.g. „edit all layers“), and which layers are selected and active within the layer stack. Only one pixel selection can be activate any time. You can save/export them as spare channels or files, and switch between them (load spare channel to pixel selection). you can’t have different selections for different layers at the same time, but you can use different selections one after the other. another way is using masks, which are „materialized“ selections, and multiple can be in use for different parts at the same time.

The „virtual“ rectangle you observe is the bounding box around the visible layer content. There are other issues documented where Affinity fails to render the canvas, and rendering artifacts occur aligned to these bounding boxes. (Mostly when dealing with alpha and having more than one vector mask, combined with alpha aware adjustments) Note that wrt bounding boxes I relate to rendering level mip map process boxes (zoom level depending), and not the actual layer bounding box (move tool). Of course, both can look identical like in this case, but not in general. To summarize: clearly a rendering bug, but I assume based on already identified common root causes (links I posted before). If you want this clarified, you may open a new bug report in the bug section, and link to this post (in the questions section).

While you are waiting, help yourself: just add these 4 adjustments on top of you file, and activate one of them to knock out the color channel. You can add them as assets for easy reuse in any document. If you need to edit channels separately: create one greyscale layer per color from CMYK, change blend mode to "add" add a "X only" channel mixer in nested position, x from CMYK These are only intended as workarounds. I fully support your request. cmyk channerl isolator no2.afdesign

This is how Affinity works, you aren’t missing anything. Best is to save edited files in a new folder, and delete the originals later (if you are sure the files are no longer needed).

Unfortunately Affinity does not support Camera Profiles currently. Other Apps support an automated workflow using Camera Profiles. In Photo, you can manually create adjustment layers (e.g. curve) to match a photography of a test page with the digital reference test file, and save these as LUT, so you can re-apply them with one click on other images. There is a small chance that forum users like @v_kyr may provide some scripts which automate that process. As an rough idea: icc profiles and curves / channels mixer adjustments are probably related in their core functionality. LUTs are simple plain text files with conversion factors to map input to output values. iCC profiles do the same, but with an binary format.

Alternative approaches: create a fully opaque backfill layer, and merge down to that layer Or create a fill layer with minimal alpha value (1/255). This will result in a an almost transparent edge, to keep the pixel layer in canvas size. You may create an levels adjustment layer which corrects the alpha values from 1 to 0 again. The handling of transparent pixels in Affinity has been discussed intensively in this forum (e.g. deleting rgb values for zero alpha, exporting from artboards vs. exported size), as it does not meet expectations of users for certain workflows, or how other Apps deal with alpha channel.

This doesn’t worked (I tried). Affinity always immediately removes transparent edges from any pixel layer (move tool / transform panel shows actual size). if you want a pixel layer of canvas size, you need to fill at least one pixel at every edge with a non-transparent color. Of course, the non-transparent pixels can be even outside the visible canvas, and the edges inside the canvas could be fully transparent. Simple merging down to an empty pixel layer does not work. But this discussion misses one important point: the result may need to be exported, and as we know, the export functions has its own will how to deal with transparent edge pixels (may surprise you).

Can you try to activate the navigator panel, so the thumbnail is visible select erase brush in the brush panel, select the basic round brush, of size 64px click the “more” button in the toolbar, inspect shape of brush Try again to erase something from the pixel layer Inspect the Navigator thumbnail if erase effect if visible. Change zoom level a bit and check canvas again if erase effect if visible. The mouse arrow changes it size in the video. Do you have any 3rd party software installed, or activated accessibility tools like loupe, zoom? Any special input device (mouse, touchpad, …)?

How do you define “expand without transforming”? filling transparent area with a color or pattern? Add fill layer below current layer. using the edge pixels repeatedly to fill the area?depends on shape (rectangular / irregular) other Intentions? Please explain.

The issue matches the older one: In Designer (or Publisher), a pixel layer nested to a layer layer or group layer causes severe delay of paint brush. In this Case, you need to zoom to get a valid rendering. You can identify the issue if you use the navigator panel: it’s thumbnail shows the correct result, only the canvas is severely delayed, until using zoom to enforce a redraw. The observation that the redraw issue is related to the (inactive) rectangular vector mask is no proof of totally new unrelated bug.

Using rectangular shapes with white fill color, nested to the masking position of of the individual images, should do the job. you need one rectangle per source image. then use transform panel / move tool (and the correct options, e.g. lock children). If you are looking for fixed frames, this thread may give you inspiration

Probably related to one of these older cases

Done. Sorry for delay. Another user raised a similar issue, so Search for older cases lead me back here.

I have some difficulties to understand the exact circumstances. single layer file? or single bitmap layer and cascade of adjustment / filter layers (everyone having their own mask layer) when you rasterize, you rasterize only the bitmap layer, but keep the other layers unchanged? There are some issues if you paste adjustments layers from other documents or assets, possibly of different resolution or dpi. Do you utilize the inherent mask of adjustment layers? Otherwise try channels panel, layer alpha, fill to get rid of unneeded inherent masks.