kirk23

What do you mean "import UV unwrap" ? Any image soft, even Microsoft Paint, could import UV unwrap if you render/ export it as image from your 3d soft. You can just print screen it from unwrap window, then paste and scale in image editor but sure there is a straight saving way in Maya too.

I also do textures and don't need any true 3d mode at all. Work on flat 2d unwrap just fine. It's lack of some conveniences and features that prevents me from using Affinity Photo/Designer . First and most important one is ability to link transforms of layers nested within different groups and mask stacks. So for example if one group is color map texture and another group is normal map one I need a way to scale/move certain layers simultaneously whenever they are placed in the stack and whatever groups or parent they are under. Another feature is compositions . A way to quickly switch and save the document to showing only specific groups and layers: color map ones. normal map ones etc. Ability to do clones/virtual copies/smart objects of certain layers within different mask stacks/clipping groups and keep them linked to each other including transform links. Ability to offset/wrap an embedded image layer ( for tileable texture ) as on/off effect and still have it embedded. Can do so in Photoshop. can't in Affinity. Krita is not yet but almost there too .

I hate dark interfaces. Never understood why they appeared.

+1 Would love to see it working within Affinity Photo mesh deform too. It's already the best of all mesh deforms available in image soft but lucks that feature badly.

While I love Affinity Photo mesh wrap still it's a lot easier to do such kind of image deforming in any 3d soft like 3d max for example. The reason is lack of control. I 'd like to be able to select several nodes and then scale and rotate them around pre-defined centre or just median centre, or first selected at least

No, what he shows in the video is basically a variant of cloning from one layer/image to another . An ancient long existed feature in many image editors. I request totally new thing not related to manual cloning but rather automatic inpainting of selected areas . I need first to do content-aware / inpainting whatever fill of selected area on one layer based on same layer content. Then a second another layer inpainting that have to repeat all the cloning/ inpainting process,pattern, scale of each cloned fragment of what I did on the first layer now on that second layer and with content of that second layer. Think about it as if you do inpainting in red channel only and then need to do same inpainting in green and blue channels separately with resulting image not turning into a mess. Kind of "repeat inpainting" in filter menu like "repeat mesh wrap" one. No other soft can do so still would be a huge step forward and competitive advantage that would change a lot the whole CG area. I have no idea if it's possible at all but perhaps it's just a kind of macro recording to be then re-executed.

Not sure I understood what you offered . Yes , I want to do it on a separate pixel layer but want to do it by inpainting/ cloning/stamping pieces of that second pixel layer in same manner/ pattern / scale as I did it initially on first pixel layer. How copy-paste would help here?

I love how I can do same mesh deform on another layer . Would like to do the same with "inpainting" For example, one layer is representing color info and a second one normal map or depth map or roughness, whatever. Now after doing inpainting on base color layer I want to do another one on the second layer matching the first one. In a word I want a feature that would keep in memory how initial inpainting was done and then do it on the next layer with 100% matching pattern

The depth info is just an obligatory image channel for any Computer Generated graphics nowadays. Simple RGB + alpha approach is so outdated. And at least to do same tricks Photoshop can AP just needs a single Photoshop feature really, an ability to lock transforms( scale /move/rotation) in between layers nested within different mask stacks and image groups . But I would rather expect something more elegant like clone layers in Krita where I couldn't care about links and the soft would do it automatically. Or rather a complete multi-channel depth aware solution with automatic depth combine in between layers. I know, there are node based tools like Nuke and Fusion to do all of this but they lack simplicity and convenience of a normal/ regular image editor so why not to do one that could cover a bit more than just simple photo retouching and web page design. I believe there is a big market niche for such image editor nobody have occupied yet. And modern gamey videocards are able to do all those necessary calculations with a speed of light.

Here it is. What I am trying to do is removing some low frequency bumpiness /errors from photogrammetry scanned piece of a ground to make it less repetitive. Somehow I like what "feature protection " does except that stepped appearance.

https://www.dropbox.com/s/tw4dyk0bxxvsdo8/DepthGrab01_AF.afphoto?dl=0 this file should work

Looks like it's something on Dropbox side currently , I can't open the same file too once I download it back from dropbox. Hardly able to download something at all I'll try again later, My problem is not the compression since I did what you offered , exported as psd. I didn't do SVG either. ??? The problem is only "feature protection"

same example as jpg

So we could could compose things with depth like in this psd example: https://www.dropbox.com/s/nmv3nunfln2w7jt/depth%20combine2.psd?dl=0 Set auto-select to "layers" in Photoshop property panel and then try to move objects. Notice how the objects auto intersect each other. Imo AP needs just a very few additions for this. I am sure it's even possible to do in much more elegant manner.

Yes, I did use psd, and actually found what's causing that kind of stepped degradation . It's "feature protection" option in frequency separation dialog. I actually love it and find pretty useful. Have no idea how to recreate it within Photoshop. Too bad it turns unusable for my tasks. Here is file example: https://www.dropbox.com/s/tw4dyk0bxxvsdo8/DepthGrab01_AF.afphoto?dl=0 Background is original tiff grabbed from zbrush . Regular high pass is more or less ok , But once I do separation with "feature protection" it turns stepped. BTW Zbrush hangs when I try to import psd output from AF. So I have to resave it within Photoshop.

I am trying to make a frequency separation on a 16 bit grayscale image from Zbrush representing the depth to get rid of some low freq bumpiness/hills from scanned piece of surface . At first it looks ok. I especially like the feature protection tolerance slider. But then, once I bring the image back to Zbrush it turns into stepped appearance as it would be regular 8 bit image. Doing same trick in Photoshop with highpass turns just ok. Should I switch something in profiles/ color rendering settings or maybe it's just a bug?

Photoshop imo, as most of other image editing tools too, is way behind of need/ features modern CG industry starves for. Its 3d painting mode is just ridiculous. Layer system is overcomplicated and quickly turns into a mess even a file author can't understand/find a thing after a while. Brushes are archaic. 16 and 32 bit modes are slow like hell. Almost every 3d soft did a huge progress during last decade , revolutionising computer graphics , and only image soft gone only through a very few cosmetic improvements. Imo we don't need a Photoshop clone , we need something new and better.

I love mesh deform tool in AP , probably best one among image editing applications. Especially love its "source" mode. It's one of hidden treasures that makes Ap so special. Yet I have to use Blender to mesh deform images. Mostly because of inconvenient node control in AP. It needs a bit more advanced one 1. ability of one click selection of a node row and then to rotate the whole row around first selected node 2. ability to scale in same way. 3. ability to relax the mesh Just look at Blender or any other 3d soft to get the idea. Also would be nice to have an option of applying same mesh deform on another image. It works on another layer but not image currently and unfortunately I can't mix 16 bit grayscale images with color ones.

IMO for CG industry any modern image editor should be capable of: a. painting within several layers representing different material properties simultaneously b. composing "depth aware" with a live real time mask encompassing pixels having higher depth value than ones beneath c. doing "live" image wrap like Krita or Painter for making repetitive textures. What would allow to do so are a kind of "virtual copy" or "clone" or whatever layers inheriting their origin layer's transforms/scaling whatever group or parent or mask stack they would be in. For point c. those clones should have a specified shift value . For point b. a layer mask should have a special stack of "depth" layers inheriting their origin's layer transforms and not the parent one . So in fact we need those clones with an option to be "chained" to each other even when inside different groups /stacks and being able to have unique blending mode and filters. For the start would be nice Photo have those photoshop styled chains/links in between layers . The ones allowing to link layer transforms whatever place on the stack those layer would be nested in

My guess Affinity is focused more on photography and graphics design . Probably those areas provide much deeper customer pool but why not to add a few extra features to make the soft useful in CG area too including movie and game productions. I was an ardent Painter user for years but it so behind of current needs nowadays. There are a few modern programs for games/movies like Substance Painter/Designer, Mari + node based composers. Still all they are inconvenient like hell imo. Over complicated for things that should be easy like 2x2. Filled with a lot routine operations and checkbox craziness to be "flexible" . I need nothing "flexible" , I need something convenient, quick and simple like hammer and sickle or a brush and palette. So I believe there is still a big market niche. In fact I need a kind of Painter ( Corel lost my hope ) capable: 1. to paint in multi layers simulationly with a special set of dabs for each physical property of a material. 2. to compose groups with depth/height combining having auto generated masks to separate pixels having higher "height" value. 3. Painter's or Krita's "wrap" mode to do repetitive textures For point 1 Affinity Photo needs just a kind of "clone" "virtual copy" layers at first + ability to lock transforms for those layers even if they are parented within different groups. Same for point 2 plus ability to have "live" masks, generated on line by a specific combination of layers . The feature is already available in Photoshop for example as "group clipping" Would be nicer if it would be just a simple command like layer A + layer B / Layer C having those layers set invisible somewhere beneath the main stack. Would beat any node based crazy Gordian knots. For point 3 it's again transform locked "virtual copy" with a specified shift value. So it's just a very few steps imo.

I have Lightroom and hardly ever open it at all. As a DAM it's absolutely awful . Imo iMatch is a best DAM on Windows. Would love to see something like this on Mac + maybe borrowed easy super quick keywording from Photo application.

Would be nicer if we had Photoshop styled layer compositions or Xara styled " name" tag . Or just color tags and a button to see only similarity "colored" layers That way we could switch multiple layers at once.

I'd love to work with 32 render passes too . Only problem Photo needs a few tools from node based Fusion and Nuke to make any advantages from that. Or at least a kind of "formula" driven layers and masks where I could use output of certain layer group as a mask for another layer . Photo ability to do that are extremely limited. Not even at Photoshop level comparing with its clipping groups feature appeared in cs6. .

Would be nice if both Photo and Design have a special "depth combine" mode for images/layers representing depth as grayscale value. So that kind of "depth" layer would mixing to underlying "depth' layers with a kind of intersecting mask, preferably antialiased. I could do so just by setting this layer as "lighter" then copy paste underlying image on the top and set it to "difference" , then threshold so anything grey turns white, then add 1pix blur to antialias the mask . So basically I would get a mask encompassing pixels with depth grayscale values higher than underlying pixels + 1 pix antialiasing. Would be nice to have all those calculations working just as a blending mode

No, I mean other thing . What I mean is called "animated brush" in Gimp, "Image hose" in Painter. "Picture tube" in Photopaint etc. In fact the only kind of brushes you need 90% of time. Not just tint variation of same black and white brush dab but rather many different/random dabs having each its own color picture/info.