dmstraker

Attached is my complete set of notes about blend modes, now spread across 52 pages of A4.
These are significantly updated from the previous revision, with lots more 'recipes' and organisation of these under main topics, hopefully to make them more usable.
Enjoy.
Blend modes notes V5.pdf

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

For the failure you experienced, I agree.
There are other failures where that is not possible, e.g. where the complete file has not been read into memory, yet, and needs to be fully read before the app would be able to Save anyplace else. And I can conceive of scenarios where that might apply even for the message you got.
Cloud services can operate in many ways, and there are things they can do that will cause problems. Some examples:
Some services may rename the local files during their processing, which will cause a problem if an application like one of the Affinity apps creates the file initially, then Opens it, then writes into it. The cloud client may perform the rename after file creation, but before the Open, causing the Open to fail. Or, similarly, the cloud client may lock the local file during their processing, and then the Affinity app may experience a failure trying to Open, or trying to write.

I've just lost hours of critical work after "Save failed because ownership of the file could not be verified. The document must now be closed."
I re-edited the image. An early save worked. A later save didn't.
To say I'm unhappy would be an understatement.
Version 2.2.1 Affinity Photo
Edit: I tried resaving the original image (a jpg) with different algorithm (Lanczos) and restarted edit. I saved at regular points along the way. Eventually I hit the same barrier. I exported the original jpg as tiff and restarted, with the same effect. I also hit the same problem with a separate image which was a RAW .NEF file.

It's a new one on me and I'm still puzzled as to what exactly might be going on. I've not heard of this happening with other applications. It would be enormously helpful if AP is unable to save remotely that it gave you an option to save locally rather than crashing out.
Nevertheless I'm grateful for the knowledge and will follow your advice, Walt. Thanks!

YouTuber InAffinity has some great tutorials explaining all the basics, especially the older ones.
 

I have the same issue. See three examples with luminosity range mask in preview mode.
Mask in the first picture is ok, but as soon as the curve touches the uppest x-axis (second picture) pure white becomes suddenly black .
I'm able to partialy workaround it moving the most right node to the left. The curve in the third picture is very close to the second one, but pure white remains as is.

Hi @dmstraker@eftee74
Many thanks for providing detail, this issue is currently logged with the developers, I've bumped the issue with your reports.

Ping! Penny drops on what rgbtoi does -- luma values. Thanks. For convenience I sometimes refer to this as 361 - RGB approx 30%, 60%, 10%. Ballpark of course, thougn on-screen i found little discernable difference when picking 1% apart. It was an aha for me to discover the difference between HSL desaturation and Vibrance desaturation.
Don't know if this is relevant to the discussion, but a very quick way to do a very flexible pixel layer mask is to:
Duplicate layer or merge visible as desired. Add new pixel layer above. Group duplicated layer and pixel layer. Change blend mode of pixel layer to Erase. Add pixels by any method to the pixel layer (including partially transparent ones). Erase blend knocks out all pixels below but stops at the end of a group. It looks like it uses the inverse of pixel alpha, so adding pixels removes ones below.
I think that in this, only the alpha value of the pixel mask layer influences the mask effect, not luminosity calculation.
 

One last note, and perhaps something that Serif can address.
In the Gradient colour panel, when you reduce opacity, the colour as displayed decreases to grey. But if the colour is still being used (as it is) in creating the gradient, then perhaps the colour should remain, hence giving some indication that you are going to still get the colour within the gradient. 
Image attached.

Understood and accepted.
There's also the psychology of perception to consider. Looking at the red-with-zero-alpha node in the gradient panel or on screen, it appears to be transparent node. Yes, showing it as 'red with zero alpha' could be tricky. But cognition is often based on simple visual cues more than an examination of further panel data. And so (as with me), it is easy to assume that what is seen is all there is.
This is what, in my development/QA/HF days we would have designated a 'usability defect'. There is clear logic in why it is what it is, but nevertheless it is experienced as a confusing difficulty.
What the solution is here, I don't know. Nor is it a hill on which I plan to expire. I'll leave it to the good people in Serif to figure what to do about it.

Attached is my complete set of notes about blend modes, now spread across 52 pages of A4.
These are significantly updated from the previous revision, with lots more 'recipes' and organisation of these under main topics, hopefully to make them more usable.
Enjoy.
Blend modes notes V5.pdf

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

Well done and thank you to everyone at Serif for a momentous 2022. Please do have a splendid Christmas and New Year break, and try not to think about work too often (difficult, I know).
As we say in Wales, Nadolig Llawen a Blwyddyn Newydd Dda (Merry Christmas and Happy New Year).

That is faulty logic, the gradient goes from black to red regardless of opacity. What happens with a gradient with the last stop with an opacity of 10%? should the gradient be shifted over so only the final 5 or 10% of the distance has any red instead of the original 50%?

Why do you rate this as bug?
the gradient will create a mix of:
black/ 100% alpha
via black 50% alpha,  red 50% alpha
to black 0% / red 0%
so for me it is correct that you see a mix of black and red with reduced opacity at the middle of the gradient.

Hi Dave,
I'll log this as an improvement for you. I don't think we've ever considered using it for anything useful but you've made a valid point. Thanks!

Being able to increment values with the mouse wheel in number boxes (often next to sliders) is wonderful, and the use of Ctrl and Shift for smaller and larger change is brilliant! 
However, only the left Ctrl works for the small increment function. The right Ctrl does nothing. 
This may seem like small beer, and for right-handed people who hold the mouse in their right hand, using the left Ctrl is natural. However, for left handed people this necessitates an awkward, cross-body stretch to perform this maneuvre. Many of the increasing number of creative people who use Affinity are lefties, which adds further to the priority of implementing what seems likely to be a trivial fix.
Thank you.
 

Attached examples. Pier1 is original. Pier2 is edited. Zoom in to see detail difference.
 
It's an experiment that gave an interesting result.
 
 

Small celebration. I've just passed the 700th video in my InAffinity YouTube channel.
There's also a web-based index, including videos sorted by time and category, plus free resources.
I'd also like to take this opportunity to pay tribute to Serif and all of its people, including devs, QA, and everyone else. You folks have created an amazing product that has freed many of us from the Adobe taxation system and indeed that stands on its own as an excellent photo editing and graphics system. Having worked in software development, quality, marketing and other areas, I have some inkling of the effort and sheer dedication that it takes. Thank you and well done.
 
 

Here's a way to make a kind of cartoon of an image.
 
Layer/Duplicate layer. Filters/Detect/Detect Edges. Layer/Invert. Blend Mode: Colour Burn. Add Curves above it all and pull up midtones. Play with Opacity to vary strength of effect. Interesting effect with landscapes and architecture. Not tried it on people.