NotMyFault

There is no way, except applying shear to a layer, adding gradient, modifying shear angle of gradient, and resetting shear of layer. The shear capability seems to be bound to the layer, not to the gradient. Saving gradient as style does not include shear.

Does duplicating the shape before expand stroke work? You can then combine both curves by merge curves or create compound, but it depends on many details. And example would help to understand your request.

Hi and welcome to the forum. I Assume you are using Affinity Photo. The easiest way is to have separate layers for body and left / right wing. Just add a motion blur nested filter to masking position to every wing, and adjust strength as required, and point of origin or angle. Other blurs may work better, e.g. radial blur or ripple.

Save them as style. https://affinity.help/designer2/en-US.lproj/pages/Panels/stylesPanel.html https://affinity.help/designer2ipad/en-US.lproj/pages/ObjectControl/styles.html

This sentence is a bit misleading. The noise settings affects any color, but for vector layers it can be changed any time after initial creation, whereas for pixel layers the color incl. noise must be chosen before using a tool like fill tool or brush. A dirty trick allows to set noise later for pixel layers: add a group, and nest the pixel layer into it now when changing the color on group layer, the pixel layer gets the chosen color assigned in real-time. but beware: using a color on a group will enforce this color to all child layers, any pixel content will be overwritten. one of the following tool must be active: hand, move, fill, crop, color picker (others might work) shape, clone, or others do block the color panel (at least on iPad) when a group layer is selected in layer stack if you have a pixel selection active, fill is restricted to the selected area.

Can you copy both layers into a new document, test again? if still reproducible, upload file. We need to see the full layer structure. Please include them as screenshots or in any screen recording, If not reproducible in new file, we need the original socument. in your case you want to snap to layer keypoints and layer bounding box only. Deactivate all other snapping options.

And this

There was a performance option in V1 called „precise clipping“, off by default. Can’t find this in the help for V2. Set it to on if you find it.

The noise filter has a bug in my view as it affects all color channels and alpha channel. If you need noise only on color channels, you can add a procedural texture filter with a noise function of your choice and activate it on RGB channels only. This will the work as intended when nested into the group into child position atop the pixel layers. noise(x,y) activate RGB

What I tried to express is you add 2 instances of the channel mixer adjustment, and adjust the settings accordingly. This allows to mimic the "math equations" of Siril. I have no more details or tutorials about the actual formulas for users coming from Siril.

Should be quite easy. use 2 channel mixer adjustments. one passes red channel only, the other passes green/blue filter only (weighted average). then use gradient map or any other method to assign colors to the greyscale images. can someone provide example images?

Not really. You need to put a larger rectangle behind the cropped image (after being cropped). PDF shows transparent background. this could also explains why different reader apps give different results, as they treat transparent areas differently. The issue is caused by semi-transparent edge pixels.

Which zoom level do you use? Do inspect noise or filtering, you must use 100% zoom level or integer multiples. For rendering on display, noise is rendered in display resolution, and when you are zoomed out the noise effect is exaggerated dramatically. Only when you merge visible or export the file, noise gets rendered in the actual document resolution. So always use 100% zoom when judging noise or sharpening.

Can you share the file? It is probably one of the adjustments or live filters (unsharp mask) which is very sensitive to zoom level. Please try a „merge visible“ in both color depths, and compare those visually at 100% zoom level. Whenever you have any noise or sharpening, or fine details with high contrast, this can happen due to mipmap rendering. Color format changes will impact areas with partial transparency, by masks or layer fx effects which include blur. You may get steep alpha value changes in areas where alpha is almost zero. This can lead to exaggerated rendering artifacts especially when zoomed out like in your case. Zoom the attached file to 100%, 200%, 50%, and any levels in between. Let Photo a few seconds when changing zoom level to reach full rendering resolution. You will spot insane rendering artifacts (false color, false lightness). Unfortunately this is by design, and requests to add features giving trustful preview where not implemented yet. rendering bug at 100% rasterized blend rangeafphoto.afphoto

Interesting, I did not spot that detail

my 2 cents without much testing: Those issues occur from anti-aliasing at the edges of the rotated image. The edge pixel becomes semi-transparent. In most cases Affinity then mixes the colors with those Fram lower layers, or black in case of transparent areas. This is invisible for high resolution, but becomes visible with very low resolution. The jpeg compression algorithm create what looks like noise. To avoid, you may need to add a white backfill rectangle (unrotated). Alternatively, a levels or curves adjustment to boost partial alpha to 100% opaque can repair. Maybe rasterise will mitigate the issue. A co-factor or trigger is color space conversion. BTW: you can spot the darker seams even in your screenshot called "original image" at the lower right edge. Inspect the images with Photo and info panel. The background is not pure white.

Forgot to mention: on M1 based Mac, Levels on alpha is broken (unfixed bug). Use curves adjustment instead, choose alpha, set left node to 0.5/0.0, set right side to 0.5 / 1.0

In cas you are using the non-destructive workflow, you can repair the clone brush behavior: add a new pixel layer activate clone brush select a square brush. Set the brush to 1px size, 100% hardness (in „more“ / brush icon, not the numeric settings, set wet edges to off) set brush source to current & below choose source position paint in clone area when done, add levels adjustment to clone layer (masking position) choose alpha channel set black level to 50% set white level to 50% now the result is hard, the former soft edged get „repaired“ to hard edges. This works reliable as the brush anti-aliasing is only applied to outer edges.

Another observation: on iPad, the clone brush has a fixed hardness setting of 80% which cannot be adjusted in the regular numeric input fields. Even when you choose a fresh square brush. You need to enter the (formerly) „more“ settings by clicking the small brush preview icon in top row.

Instead of using the clone brush, you may use alternative methods: make a gross selection of the rough area (larger) copy/paste the area have snapping on with fore pixel alignment on, move by whole pixels off, other snapping off add a mask layer to clone layer use pixel brush to paint in mask or remove unwanted areas Check that the mask is 100% hard (isolation mode)

You are welcome. We all have moments not recognizing the obvious

Layer anti-alias settings impact vector layers only, and have no impact to pixel layers. Pixel based brushes and tools always have forced anti-aliasing. You cannot deactivate it. So when you use pixel brushes, you will get multiple levels of antialiasing and resampling atop of each other: pixel brush anti-aliasing for every brush stroke inside any bitmap layer Vector layers have anti-aliasing settings in blend options resampling settings (in performance settings for in-canvas rendering, in export settings for export, enforced with no method to avoid when using rasterize) like bilinear for rendering and export

Brush flow is zero.

Yes. The steps are a bit different to the PS based video, and GarryP showed them in his video. add pixel layer (or rectangular shape) in shadow color e.g. dark blue/gray. set layer blend mode to multiply nest the shape to the main layer in child position. It will be clipped automatically, no need to use a mask from base layer. add a black mask via menu, and nest it to masking position of the shadow layer. use paintbrush to paint in white over areas where you want to apply the shadow It does.

All said sounds more like a generic „brush does not work“ issue not related to RGB/32. Whenever you observer a brush not working, the first things to do: save file with history. This really helps to analyze the issue. Do not close Photo, let it open Create a new document, add a pixel layer, try the same tool (pixel brush, erase brush) there. Does it work? Use layer isolation /solo mode use navigator panel and make it really large make a screen recording using the Windows default tools.