Bryan Rieger

Thanks. Hopefully, it is addressed. It's gone past annoying that I must flip that stupid toggle every time I want to create a new document from a preset that isn't in the same portrait or landscape aspect ratio as the previous image I created.

Hi All,
We have just made 2.2 live across all App Stores and our website. As always, the feedback we have received from you all during this beta has been an incredible help in both shaping the new features as well as squashing many bugs!
The build we have released is the same as RC3 (build 2005), meaning for the time being whether you use the latest beta version or switch to using the release version they are the same. If you anticipate taking part in the next beta you may wish to keep your current beta version installed, as this will be updated on auto update when a new beta is available. This will either be a 2.2 patch if we feel one is needed, or moving straight onto a 2.3 beta.
It would be impossible for us to do what we do without your support, and I just wanted to say a huge thanks for everything you have contributed to this release.
All the best,
Ash

@CM0 thanks for sharing this, as I'm often left wondering how and why layer FX render as they do. While this doesn't solve the problems, it does help develop an understanding of what works, what doesn't and how to approach using layer FX.

Using Affinity for several years and spending hours in it per day, I have been constantly troubled by inexplainable behaviors that when reported never seem to have adequate explanation. Is it a bug or working as intended but in some bizarre way?
Well, I just spent many hours over the weekend finally attempting to map out in detail how everything is rendered in hopes to bringing some sanity to my understanding. The exercise has been beneficial in that I can now finally explain many of the oddities; however, I can't confirm whether they are intended, but can point to why they occur.
I hope that Affinity could confirm/document the true rending order. This is very important as it helps us to understand how to properly create compositions as well as sort out bugs from simply lack of understanding. The exercise of documenting would also likely bring to light some of the non-intuitive behaviors such that hopefully they could be addressed.
It would be great if the rendering were simple, something such as child layers > current layer, FX, blend ranges etc. However, what I discovered is that it is anything but simple.
This is my best attempt so far, it may not be perfectly accurate, but has been good enough to explain all of the odd behavior I have seen.
Rendering pipeline order and rules
Child layers of group
(if passthrough) bottom of stack is rendered on top of layer below group layer
rasterizing this layer may result in a different blend.
use merge down or merge visible to retain original on screen view.
(else all other color blend modes) The bottom stack starts as if on top of transparent.
this is why for the bottom layer of the group, its blend mode will usually have no effect and blend ranges for underlying will have no effect as there is nothing to blend with.
Transparency tool effect and Masks
Adjustment layers | Live Effect layers
Layer FX ( rendered to temporary FX layer )
Exception - Gaussian blur will result in immediate merge down as well as blend range underlying being applied to source instead.
Exception - when combined with of the following then will be applied later after underlying blend ranges
Other layer FX
Preserve Alpha
Current layer is any blend mode other than ‘normal’
Exception - Child layers
Effects are applied to the child layers after child step 4. Or in other words, applied after layer FX merge down. So blend ranges underlying blend is not utilized from children when layer FX are rendered. The fully rendered result of children is not used!
Blend Ranges ( source blend )
Exception - Gaussian blur disables source blend unless combined with another FX or preserve alpha.
Exception - ignored if Live layer is applied to current layer and current layer is an image layer.
Layer FX previous rendering from step 2 merged down
Blend Ranges ( underlying blend )
Exception - This operation skipped when Gaussian blur enabled unless combined with another FX, preserve alpha or blend mode other than ‘normal’
Exception - ignored if Live layer is applied to current layer and current layer is an image layer.
Layer FX ( Gaussian blur exception )
Gaussian blur applied at this step when combined with other FX, preserve alpha or blend mode other than ‘normal’
Merge using blend mode (Multiply, Screen etc)
I have found additional oddities, but still can't sort out the rules yet. For example Bevel, outline, outer glow and outer shadow all render in a different way than the other FX in certain conditions, but can't yet articulate the reason.
These are some major notable observations from the above:
Gaussian blur seems to be the greatest offender of unexpected behavior. So much so it actually changes the rendering order and will completely flip the blend ranges depending on certain conditions. Rendering does not seem to be serial. It is interwoven or forked. See step 4 layer FX. This adds substantial confusion when trying to figure out what is the right order for an effect you imagine. It also makes some combinations impossible. Live FX completely breaks blend ranges applied on a pixel layer. I have refrained from specifically listing anything as a bug although I doubt anyone finds the above as desirable. What I hope is that we could finally officially document what is supposed to happen and from there figure out how to improve this to something that works as desired and is intuitive whether we call it bugs or enhancements we just need solutions.
I make a lot of complex compositions and the above always costs me hours of work attempting work arounds for unexpected behaviors.
Attached is a test file I used to evaluate much of the above. The key parts is that it has blend ranges for both source and underlying applied in a way that is easily recognizable. There is a character '1' on a blue gradient background and it is blended such that the top of the one is transparent for the source blend ranges and the bottom is transparent due to the bottom blend ranges. Experiment with applying FX directly to the image or parent group in different combinations.
 
Render order testing.afpub

And it is the better case 🙂

I’ve yet to see Serif comment on (or even acknowledge) any potential memory leak posts which doesn’t fill me with confidence. The Affinity 2.x apps memory usage behaviour doesn’t resemble any other application I use (macOS) in that they tend to hold on to memory, grabbing more with every new document), and never fully releasing it unless the app itself is closed (or crashes, which is sadly not uncommon in v2). Whether this is by design or not has never been addressed by Serif, and having another random user ramble on about ‘how memory works’ is getting old as only the Affinity apps exhibit this behaviour.

Perhaps a minor for some people, but it sure would be nice if .afbook files used relative paths for chapter references. Rarely, if ever, have I had use for an absolute file reference.
It makes switching from laptop to desktop cumbersome when using cloud storage, as well as with any collaborative workflow.
(If I'm missing a setting for using relative paths, I'd love to hear about it  ).

@walt.farrell in presentation/kiosk mode without any app chrome (toolbars, etc) and/or without a keyboard (for navigation) having navigation in the document itself is really useful.

Using anchors is a lot of work. What would be nice is simply being able to drop a couple of 'buttons' on a master page and assign them the behaviours of 'next' page and 'previous' page (in sequential order). As the buttons are on a master page they would be available on every page that master is applied to, with the final hyperlinks being applied in context of the actual sequential pages in the document on export.

The 'swatches' button at the bottom which reveals the swatches panel is really unintuitive, as it slides over/replaces the whole colour panel. Why not simply extend the colour panel vertically and display the swatches underneath as it's only possible to have one panel open at a time anyway. It can't be about vertical space as the Pixbay (why?) panel extends the full height, as do the adjustments panel, history panel, etc). The size of the swatches could easily be reduced in the swatch palette, and the 'Swatches' button could easily be repurposed for the swatch group selection.

Wow, @Aammppaa that works great. Not intuitive (especially the whole symbol sync thing), but it's exactly what I need. Thank both @Aammppaa and @retrograde for your help with this. I've got a few ways to workaround this issue now.

Wow, @Aammppaa that works great. Not intuitive (especially the whole symbol sync thing), but it's exactly what I need. Thank both @Aammppaa and @retrograde for your help with this. I've got a few ways to workaround this issue now.

Good solution @Aammppaa I removed the heavy stroke on the shadow (bottom) layer and added a drop shadow with an offset to the bottom right for an angled shadow. Works great!

Use Symbols.
Create your sticker as a Symbol. Duplicate underneath. Ensure Sync is OFF, and add a white stroke to outline everything. Duplicate the sticker again. Sync still OFF, and fill black, set opacity, add Gaussian Blur FX if desired to get shadow. Addition of a black stroke allows control of Shadow size. Turn Sync ON so that alterations to the sticker propagate to the Outline and Shadow versions.  
 
Sticker with Outline and Shadow.afdesign

Thanks @retrograde that works really well and does save me time. I do find with more complex groups the contour struggles a bit, creating gaps where there shouldn't be, etc but that having a combined group contour (with shadow FX) along with being able to select the individual elements in the layers panel is a great (if not yet another) workaround.

Not sure if this helps but... I tried using some copied combined shapes as a compound shape by marquee selecting all of the copied shapes and control clicking (mac) on the 'add' boolean icon to create a non destructive compound shape of all of the selected shapes, then you can apply an outline and drop shadow to the compound shape. It's not perfect but it keeps all of the original shapes intact incase you need to reposition or transform or remove anything later as opposed to creating a destructive 'add' boolean.
To edit the individual shapes you'll need to select them in the layers panel if you want to transform them as once made into a compound shapes you won't be able to see the individual shapes outline even in outline view mode. Then you can make a style of the outline and drop shadow of the compound shape for a quick one click application to any new shapes. Like I say it's not perfect but it's a little less destructive than combining all of the shapes into one vector shape. It might add up to a lot of shapes though if you have a ton of objects in your file.

Please reconsider providing architecture specific builds for macOS Intel and Apple Silicon directly from Affinity. I know that doesn't help with the App Store Universal builds issue, but at least you could help those customers who have purchased directly from you. Oh, and an added benefit would be showing some respect to users with smaller hard-drives by not shipping GBs of libs they don't need.
Simply saying 'sorry for the inconvenience, but we're not going to bother to at least fix it where we can' just further demonstrates Serif's contempt for its users. 

Working with the pencil tool in designer can be maddening enough at times (extra nodes/data, destructive delete, etc) but trying to set what colour you're drawing with can drive one to madness. The brush tool also has a few oddities, but it's at least a bit more consistent. I've attached two videos showing Designer on both macOS and iPadOS and how picking a colour is like rolling the dice, let alone trying to determine if the shape is going to be filled even if there is no fill colour selected.
One thing I did figure out is that the pencil tool stroke is only taken from the stroke colour well in the colour panel, while the brush stroke colour can be either the fill colour or the stroke colour. Also, the colours displayed on the context toolbar are not always in sync with the colours in the colour panel colour wells.
One thing that would help in the context toolbar is that if a colour is ONLY applied to the stroke use the same colour well (the stroke, unfilled circle one) in the context toolbar—as it is it feels like you can use either or the fill. Also, I'm not sure how Designer is figuring that if I've currently selected a line drawn with blue, that the next line I want to draw should be yellow or whatever colour it decides.
A little consistency (and sanity) would be most welcome.
Addendum: I also use the colour panel colour wells rather than the context toolbars ones primarily because a) I always associate setting and selecting colours with the colour panel, b) it has additional controls to set the opacity and noise settings, and c) I primarily toggle between the colour panel, stroke panel, and layers panel all of which are easily accessible from the top right or the screen, rather than having to find and manipulate settings in the context toolbar (which I honestly don't find terribly useful, can't remember what it contains and when, and is generally very confusing).

macos.mp4 ipados.mp4

All we’re on the internal drive in my iCloud synced ~/Documents folder.
Unfortunately I don’t recall the colour of the memory pressure gauge.

I'm visiting a shrink tomorrow. This thread is so surreal it must be a halucination.

Unfortunately I didn’t get the memo. I can only assume this magical event might be on the 3rd Monday of July (if it’s not raining and all the appropriate paperwork has been filed), as it not only did not release any memory, but continued to allocate itself more memory to the point of eventually impacting app and system performance.
 

Another try at drawing in Photo in the iPad and I think it's going better. Thanks to the staff I was able to get the blur brush working again and though subtle is did help with some blending issues. This is definitely a case where not good photo can make a much more interesting illustration. Isolating the marcher form everyone else made for a much better composition. 

The expectation of a user should be that the clipboard is NOT cleared until the user replaces its content with something else (or the user logs out or the computer is rebooted, etc.).  It should not be a standard practice of most applications to clear the clipboard themselves (password managers being a notable exception).
I have seen a few Micro$oft apps work around this by prompting the user that there "is a large amount of data in the clipboard" and asking if it should be cleared when the user attempts to quit (exit) the application.  Annoying.
Clearing the clipboard on exit could be offered as a preference, but should clearly not be a default behavior.

Some comparison numbers from my laptop Windows 11 system, for Designer 2.1.1:
With an empty Assets category selected before I last closed Designer, and an empty Brush category selected (both vector and raster), when I start Designer 2.1.1 it takes 1,782 MB as reported by Task Manager.

On the other hand, if I close Designer while I have a large Assets category selected, and a large Brush category selected, when I start Designer 2.1.1 again it takes 2,566 MB as reported by Task Manager.

Conclusion: The active (last selected) panel categories have a strong influence on the storage usage, even with no documents open. In my case, for those chosen categories in Assets and Brushes, it takes 784 MB of storage at all times.
  If I then create an empty document (8.5 x 11 inches, 300 DPI), the memory usage grows to 3,227 MB.

Conclusion: that empty document took another 661 MB.
  If I close that document, the memory usage drops down to 3,050 MB. So it did not release all the storage, having kept 484 MB of it.
  If I create another empty document (same size), I can see that the memory utilization initially drops by about 100 MB, and then finishes at 3,612 MB.

Conclusion: Assuming the new document also took 661 MB, the starting point just before it was created was 2,951 MB. And 2,951 - 2566 gives a delta of 385 MB that remains unaccounted for. This could indicate some kind of leak, but it would take more detailed analysis to know if it is indeed lost or being used for something that I just don't yet understand.

On launch: 648MB, after creating a 1080 x 1080 px artboard:  1.4BG, after creating another 1080 x 1080px artboard: 1.85GB.
After closing one artboard: 1.66GB, after closing the remaining artboard: 1.38GB.
These numbers are consistent with what I've reported earlier in this thread in that the memory isn't being fully released when documents are closed. It becomes much more noticeable after you've opened and closed MANY documents.
Addendum: after sitting minimized and idle with nothing open for 30 minutes, and allowing Adobe apps to monopolize the available physical memory, Designer's memory usage did drop from 1.38GB to 1.37GB.
Note: I also don't use Assets, Styles, Symbols, etc,—in fact I delete all the built-in ones—but I do have a lot of pixel brushes that weren't available at the time (as I was in the 'Designer' persona).