Frank Jonen

32-Bit mode - Lens Distortion has the same error the blurs used to have. No idea how long that's there, I rarely use lens distortions. To verify: Open any 32-Bit file, add Lens Distortion, nest it to a group and the blocks will appear. Take it out of the group and it looks normal again. Only seems to happen with positive values in Lens Distortion. Negative values look normal. General Bug (always been there) Layer blend options: Set two channels to zero, the layer becomes invisible while at 100% opacity. What I'd expect: I set G and B to zero, now only a view to the red channel should be available at full opacity.

It's how Brad explained Nexus.

A lot of things changed since then. It think it was around the time of CS 3 when the transition to Windows first, translate later began. Today Thomas Knoll merely works on the Camera Raw segment of the app. Now you can open up the bundle of any major Adobe app and see Windows files all over the place along with runtime converters. I don't think that helps with performance on the macOS side of things.

What would be accurate then? You stopped your reply too soon.

When you compare this to Photoshop on Windows / Mac, keep in mind that Photoshop hasn't been a native Mac app for years. Open the bundle files and you'll see a bunch of .EXE files and .DLLs and runtime tools, on the Mac version. This suggests, when you first have to run translation code as a basis for your app, it slows things down. That's why Photoshop is faster on Windows. It's the origin platform it is developed for. For Affinity apps, macOS offers developers a bunch of native hooks that go way deep into the system and hardware accelerations are based on OpenCL for the most part. This explains why it's currently faster on macOS (the origin platform here). If I got that right, Affinity's foundation seems to work like The Foundry's NEXUS platform, which is native to each operating system they support, Modo then runs on NEXUS with a few modifications for platform differences. Hopefully someone from Serif can weigh in on the issue.

Does someone know how to implement Floyd-Steinberg or Jarvis, Judice, Ninke dithering with equations? Is that even possible at this point?

I just use the gradient on the live layer of the tool I want to apply. Keeps everything live and I can iterate much quicker. The white boxes you see when adding a live filter or adjustment layer, those carry the masks, just click there and paint away. Black is nothing, white is 'apply tool' (0 to 1). Or to completely open it up. Put a bunch of adjustments into a layer group (just the filters, no pixels) and give that group a mask. You can get really precise that way. Do one overall gradient for example and in each layer add or take away where the effect is too strong or bothers you. Just a word of warning, stacking blur and sharpen filters together can lead to longer processing times when you export the final image.

The addition of PPI would make sense in a context like the persona I proposed a while ago: https://forum.affinity.serif.com/index.php?/topic/30454-professional-pre-press-persona-please/ It's fairly technical as well as practical. Not something every designer deals with all the time. Most people create assets and send them off to somebody else to deal with. Some forums are even filled with people asking what this DPI thing is when we have pixels on a screen. In general I noticed people use DPI and pixel count in a very specific context today which is when looking at a physical space "I'm filling an area about yea big with this much pixels, it'll be looked at from about that distance, so I'll need to put the pixels this close together / need x amount of pixels per inch on surface". Which is highly simplified but has been working for ages now. I think having that option in a very print specific context where people who go there for a reason won't be confused by this. It could also help to and be used for explaining the concept to others. PPI and DPI sliders for example to see how your output performs.

Here's hoping for a Substance Plug-in that translates what can be translated to Affinity native features and have the rest covered with custom controls.

One way to surpass InDesign is typography. InDesign has been struggling with that for years now and has every fallen behind Quark so badly publishers are saving money in page count when printing books by using Quark vs InDesign. But Publisher doesn't have to beat any of those. It has to beat LaTex, or at least get to parity. For web typography, supporting the latest OpenType features is a must now that the problem with legacy browsers is fading away. Colour and SVG based OpenType fonts is one thing. Variable fonts is quite another. It saves download time for the user since you only have to serve one font to provide all the weights. Recent Safari versions handle this well. Also for ePub, support of iBook packages would be great as well as the ability to create them as well as some way of easy round tripping with Tumult's Hype.

Nope, not Affinity in this case. You sounded like you needed a working LUT, so I made one :) I'm a magician, I pulled it out of my… I use Lattice for technical LUTs like this. As for the bug: Invert generates the same output as switching black and white points, that's why I tried both, to check whether it's specific to that layer. It's the export LUT function. AP used to have an option to compensate for this when loading a LUT, used to. If it still were there you could've just checked it and used the LUT it exported. If you shoot more film in the future, it's definitely worth it to invest in a ColorChecker, that way you can match the colours to it after inverting to get accurate colours. Usually it's enough to do one calibration shot per roll unless the temperatures drastically change. It'd be great if Affinity would support colour targets though. Infer LUT generally does a good enough job, so there could be a "LUT from Target" option in there as well.

OK so I got curious :) Maybe this might help you batch this quicker. I don't know what film stock you used nor do I have any negative profiles at the moment. Is this somewhat close? For debuggig purposes, I also see the bug. It happens when you use the invert layer as well as when you use levels and swap black and white point. Export a LUT from that and the 3D read order gets swapped. Inverse + Balance.cube.zip

I tried your image with my inversion LUT and it came out blue just like regular inversion. Attached you a working inverter LUT so you can keep busy over the holidays :D inverter.cube.zip

Currently the vector scope isn't scaleable, it's quite small. It's been my screwup meter for years and in Photo it's just damn hard right now to see things close to the centre because it's so small. While at it. Once the scopes are bigger, they could be used to mask colour ranges. Just draw a selection on the scope with the selection brush (or regular brush) and the colours within their ranges are masked.

It's things like this that'll move magazine designers to Affinity Publisher. Creating 3D rating graphs was one of my banes back then.

Here's the rundown on Family Sharing as it is today, from: Section D of http://www.apple.com/legal/internet-services/itunes/us/terms.html This is the only restriction left. The "Same Household" stuff seems gone completely. "Family Sharing Rules: You can only belong to one Family at a time, and may join any Family no more than twice per year. You can change the Apple ID you associate with a Family no more than once every 90 days. All Family members must share the same Home Country. Not all Content, including In-App Purchases, subscriptions, and some previously acquired Apps, are eligible for Family Sharing."

Family sharing in the AppStore preferences. You share your AppStore account with a family member (within the same household) and they'll have access to all the apps you bought that allow for family sharing.

Is there a way to see how big a file is in memory to see why AP is becoming Beachballhalla? In the info panel I only see Memory pressure which I assume denotes read/write velocity and Memory efficiency which I have no idea what it means. Especially since it displays as over 3k% which is interesting given what per cent means. I haven't found a way yet to display the internal memory use. I can sort of guess it with iStat menus and TOP but that's the whole app, not just my file. I can't optimise the app but I can optimise the file for performance once I know what's causing what to bloat.

Thanks, that works too. Odd though since the shortcut values were empty. Now I have two options to toggle them. I put them on the right and left side buttons of my Wacom tablet btw. So depending on how I work I can stay adjustment focused (hands more on keys) or paint focused (tablet mostly).

I spent most of the day working with spare channels today. Something that's always been a little bit cumbersome in a layer based app. Since you do live previews of the R,G,B,A channels this also slows it down noticeably when I'm going into the Channels tab (that's why I leave it tucked in behind the Layers panel). A dedicated Spare Channels panel would be great (tiled view, list view) to quickly swipe the pen over, right click and load a selection, or save a selection when right clicking on an empty part of the panel. A bit like the Styles or Swatches panel. Normal clicking and dragging would allow moving the layers around to organise them for easier findability. It's really a workflow enhancement to have the luxury of seeing these separately with bigger previews while leaving the regular Channels panel alone.

Exactly. :) And it prompted me to look around the misc section a bit more which helped getting my font related panels organised (all on the left).

Now that I looked into shortcuts more thanks to the comments here I found an interim solution to this. I set option+L, option+R to toggle studio left and studio right (the bigass panels) and put in the tools that I don't constantly need. That helps with painting as I still have my floating stuff and only that. Especially useful in full-screen. It's OK but I still miss being able to collapse stuff.

Alt+b does nothing but produce an error sound. Had to set the shortcut manually. It doesn't toggle back to the previous colour though, it's a one way train like in Photoshop.

32f to 16iBit output bug still present. It's noticeably less but it's still there. Also tested in easy steps without export. Flatten layer stack Transform to 16-Bit White breakup happens immediately no output file needed. What seems to cause it is another 32f layer (cg, placed and rasterised) set to Soft Light (at 80%). Once I disable that, no white breakup. The layer is masked and with RC5 I now get black dots in 32f mode where the mask tapers off. ONLY in Soft Light. just went through the other blending modes, they seem fine.

Can you add a little thing still that'd save a lot of time for 1.5? When painting in a mask layer, temporarily switch to full black (bg) and full white (fg) in the colour picker? Right now it stays colour and you're painting in the mask layer / layer mask with the greyscale equivalent of the foreground colour. Will test for the 32f white flashing later tonight/tomorrow.