NinthDesertDude

Nah, using a .bat is harder than clicking browse tbh

You can work around this probably if you right click paint.net -> Run as Administrator

You can get the pixels for translation based on the position info it gives in the toolbar, though you need to do the mental arithmetic. You can mostly get the rotation with that info too (it says the angle). The scaling is the hardest to compute, you might need to pull out a calculator to deal with that. Paint.net is a useful tool but it's not intended to be like CAD or Blender with exact translations. I think that's probably out of scope (I can't speak on behalf of course)

If there's a real resolution issue, I bet it'll go away if you use press F8 twice to close and reopen the color window. (The F5, F6, F7, F8 keys correspond to the floating tool windows.)

What tricks come to mind include: - only drawing the visible portion of the canvas (after transformations) - keeping a merged copy in memory of the visible portion of the canvas for all layers above or below the current, I don't know if PDN does this but it would be faster than merging each pixel per layer on every frame to redraw it. So this means 3 images in total: one for all layers above (merged), one for the active layer, one for all layers below (merged). For large images, you'd probably want to keep the temporary copies saved to disk or only generated for the visible portion of the canvas to avoid overloading RAM. I believe paint.net layers automatically have the ability to save & load from disk already, so it's probably doing that much at least - load and unload the image in chunks to handle extremely large images. Loading it all in memory would quickly fail for a 10x10k image - use GPU not CPU. Older version of paint.net were not GPU optimized I'm sure Rick can pour over it in more detail and with better accuracy since I'm not actually a paint.net dev. I just have a plugin.

Grab a copy of the default Steve skin and just replace all the pixels that you didn't intend to change with that. Basically put it on a layer below your modified one, in the same location, then select every pixel you intentionally changed that you know of, invert the selection, press delete, and merge the two layers. It should be faster than trying to figure out changed pixels. Alternatively, an easy way to tell what's different is to put the Steve skin in the same spot on a layer above and set the layer's blending method to XOR. That will show all differences in pixels. I also have a plugin somewhere called Image Difference that basically did the same thing.

Built-in channel view isn't a planned feature right now afaik. Lots of stuff to work on, almost no devs. The last major forays have been GPU and the basis for custom brushes. Layer groups, masks, channels, etc. are all sort of related features and maybe when the work switches to focus on that, then it'll receive attention and thought

It would be faster if there was a shortcut to switch between images you have open besides using the mouse to click on their thumbnails (if there is already, I apologize). I often use paint.net to open multiple camera pics of the same subject to edit the best one -- that's often up to 20 images and I need to flip through them to figure out which has the best lighting and is the least blurry. I can't think of too many other usecases, but I've been doing this long enough it's worth asking, I think. Low priority, but if it's also low-hanging fruit then it might be worth bringing up

Something like Just follow some basic animation principles - skip frames to make it look more sudden, which is why the start and end have only one frame. The sword swings fast and stops suddenly - add frames where you want the attention to linger, which is why there's 2 long swish frames (add more to smooth it out if it needs to last longer) I drew a thin swish above to make it look sharper. A wide swish makes it look more powerful. I'd give a sword a thin swish, an axe a wide one, but it's all about what you want Spend 24 mins sometime to learn the principles of animation, it's like an entire course in under half an hour and it's fun to watch. Not just generic advice either

Just hasn't been considered and prioritized yet due to time, I think. A blending mode might actually be possible now, depends on the code. From what I've seen, Rick prefers making unified changes, so probably no layer mask without layer groups or whatever other highly layer-related features belong together. I hear you, though. Alpha masks are a common and commonly-used feature in a lot of similar software.

Guessing a bug with Add Noise now that it runs on GPU. That or how PDN chunks out an image for parallelization, but that might've changed with the GPU stuff. What's the exact width in pixels of those "bars" in the messed up image? If it's a power of two it might be important. If you save the image, does it still look like that when you reopen it? (Probably should.) What were the settings you ran for Add Noise, and can you recreate this bug after freshly starting paint.net? Being able to do that helps people try to recreate it Thanks

Thanks for the update. Idle thought, should we make it a habit to say on release threads, since they're dependable for each version and the PDN documentation strives to be up to date, when it's been updated to mention all changes in the release (in this case, the new keyboard shortcuts)? That convention would probably help avoid anything slipping through.

Whoops, I forgot my implementation was a bit weak in that area. It does copy the pixels over but it only works as long as at least one pixel of the brush is touching the canvas on the edge you draw on, so if you keep moving the brush until it's fully off canvas, it won't wrap that around. I should change the algorithm to handle that because while it's technically seamless, it's not as easy to use

Yes, my plugin Dynamic Draw supports this under Options -> Seamless Drawing when you open it. This will seamlessly draw across the whole image, with one caveat that if you use a brush so large that it would wrap around more than once on a single brush stroke (so basically so big it's at least double the canvas size), it will stop wrapping around after the first time. You probably don't have to worry about that at all since it's an unusual thing to do. Note this plugin is suited more for drawing than geometry tools, so e.g. line, circle, etc. are not available in the plugin yet. Hope it works for you!

I've had this issue once on an older version of Paint.net; I solved it by being on a different computer. I think it might be my laptop that does that? I don't use it much though. In the meantime, Ctrl+N and Ctrl+V Edit: oh I remember what it was. It was key interception, and it was by John's Background Switcher. It has a shortcut for Ctrl+Alt+V.

visual glitch with the position of the handle when you use the scroll wheel to increase zoom to 10,000 only when using the scroll wheel for it

This might benefit from using LockBits with Parallel.For, but that depends on if accessing bit data with an instance of Surface locks and unlocks for read access each time. If so, then LockBits would perform much much better. Parallel.For probably beats making 16 tasks manually since it takes into account what the user's number of cores are, as long as you use it carefully with LockBits to avoid conflicts

I start to wonder if having gamma correction in the toolbar is going to lead to the path of having too many nuance choices available, as more functionality gets added. For now it's fine, it fits, but that's my only wonder. It would fit just as well to have it only in settings (and probably be an all-up setting that affects all relevant ones at once, unless there's a need to turn it on and off per tool).

There are sometimes good reasons to have a selection set which covers all the visible image, or just the whole image. When a selection is selected, it adds a blue tint to everything. While the layer window is open, the user is editing properties of entire layers, so there is no need to show a selection. It gets in the way a bit if you're trying to balance colors. Here's a contrived example: In real usecases, I usually have an image and a filtered copy of that image, and I want to tone down the filtered copy by reducing its opacity before merging down. In all cases though, showing selection while the layer properties dialog is open has only cons. Would it be possible to hide the selection in this case?

> Im thinking things like 10*50... Why not type in 500? easy, 47 * 23. Annoying enough that a calculator is much easier. If you struggle to find something, I did basic math parsing once with infix notation in https://github.com/NinthDesertDude/SharpParser which does allow simple functions, but also has a version that doesn't. Maybe it's useful

The bottom also shows the bounding rectangle size, which I use sometimes to manually flip horizontally/vertically (luckily there are plugins for that):

I agree, and thought about that too, though I think that was actually requested here once before and rejected as too much work. Or something similar

So after ~1 decade of manually multiplying the width or height of an image based on the currently-open image's width or height (which is prepopulated in the width/height boxes), I just realized it'd be nice to have this by default. By long and far the most common use case is for having side-by-side variations of an image (e.g. a "before and after" layout or "step 1, 2, 3"). For this, multiplying the width by 200% or 300% of the current width would suffice. I've been doing that calculation by hand for a while, now. So my request is if the UI for the New image dialog to allow the new image's width and height to be expressed as a percentage of the currently-open image's width and height. Ideally if there is no image open, these controls would be grayed out (or hidden, as appropriate). There are 2 sections right now, "Pixel size" and "Print size"; maybe a third section like "Relative size to current". Thoughts?

I haven't dealt with stabilizers yet. iirc I wrote some simple tests that weren't adequate back in 2020 and just got sidetracked by other things to do. It's something I really ought to get back to, still on my list. There's a bunch of things I need to support with PDN 5 too in order for the plugin to work without Rick's code that stops it from breaking. It's been half a year-ish since I started writing Dynamic Draw v4, which has a laundry list of improvements, but I left off in some tricky math with recreating clone stamp. I really need to finish that

No way to rotate by pen, but you don't have to rotate by slider because you can use the shortcuts: Shift + mouse wheel or Shift + arrow keys