MJW

The algebra appears to be correct. As far as the image being too white, could it possibly be that in the image-blended-with-solid-color case, you aren't taking into account that for the reverse process to work, the image must be able to be produced by blending some image with the solid color. For example, if the solid color is pure red, and the composite image contains pure green, there's no way to produce the composite image unless the alpha for the red layer is 0. The process is, I think, easier to analyze if you think of the lower layer as the color layer, and the upper layer as the image, so that it's the image-layer alpha that must be determined, along with the image-layer color. Otherwise, you're computing the color for the lower layer and the alpha for the upper layer. What you're doing is essentially what my Color Clearer plugin does. I wrote a little article on doing the math exactly with integers. Because of the integer math, it's a bit confusing, but the basic idea is that an alpha is computed for each color component, then the image color is computed based on the minimum of the three alphas. In your follow up comment, you have just substituted opacity for alpha, where opacity = alpha/255.

As much as I recommend knowing about and using layers, you can also use the Paste From Clipboard plugin to directly blend the image on the clipboard into a layer. Just enable the Blend with Canvas Image option.

I realize this would be less convenient than a plugin that does exactly what you requested, but you could write a plugin to copy the desired image to the clipboard. You could then paste the image with a Ctrl+V, selecting the Expand Canvas option (which only requires pressing Enter). That might be a easier than repeating the steps for each image. Once it was done for one image, it could be repeated for other images with Ctrl+F (Repeat last Effect), Ctrl+V (Paste), Enter (Expand canvas), Ctrt+D (Deselect), Ctrl+S (Save).

Could you please change the title to something like, "Could someone write a plugin for TIM file format?" "Can someone do this?" is a generic title, and generic tiles aren't allowed (Rule 6).

So we can mark are own replies as solutions? Somehow that doesn't seem quite right.

It would help if you'd describe the functionality you're looking for. Most of us have no clue about what exactly NCH PhotoPad's "Collage Creator" does, and I, at least, am far too lazy to find out. (Okay, we do have a clue -- it somehow helps create collages; but that's not really enough to point you in the right direction.)

BoltBait's plugin pack.

Try using Color Clearer with the Color to Make Transparent set to white. If the overall result is too transparent, you can use Red ochre's Alphathreshold to adjust the transparency range, or for more complete control (with more complexity), pyrochild's Curves+. It would be better to start with a map that goes the full range from black to white, then after running Color Clearer, run Brightness/Contrast to lighten the result to the desired range of gray. Otherwise, when you modify the transparency with Alphathreshold or Curves+, the range of transparency may be overly discontinuous. (This method assumes you want pixels to smoothly increase in transparency as the lightness increases, though adjusting the alpha channel afterward with Curves+ could allow a quite sudden transition, if that's your goal.) To change the map from white to brown, I'd try Ed Harvey's Color Filter.

There are at least two plugins that may help: dpy's AA's Assistant and pyjo's Basic Anitalias. However, both work only with "objects," which means groups of non-transparent pixels on a transparent background. So you'll need to make the background (water) transparent rather than the blue-gray color. That should be pretty easy for this example, since the color is pretty solid. Use the Color Picker to sample the grayish background color. Add a new layer beneath the map layer and fill it with the background color. Disable this layer so that the next steps are clearer. Use the Magic Wand in Global mode to select the background color. Adjust the tolerance to select all the background pixels. Delete the selection. Run one or the other (or both) the antialias plugins. Enable the colored background. EDIT: I don't know if you're showing a magnified version, but if not, each block is about ten pixels in size, and the antialiasing plugins aren't going to help with that. You'll first need to resize the image down so that the blocks are pixel in size. In fact, if your pixels are that large, resizing to one pixel per block, then resizing back to the original size will probably smooth things out provided you use the default resize method. EDIT 2: I really need to read more thoroughly, since I completely overlooked the explanation at the end of the comment that the image already consists of separate layers, and that it's magnified. So assuming the blocks before magnification represent single pixels, you can try running the antialiasing plugins on the non-sea layers.

Though in any case I'm not aware of any plugins that might help, what is the format of the image you want to convert to a Winkel Tripel projection? Unless you have an image in some other format, such as a map using a different projection, I can't see exactly what you intend to do. If you don't already have a map in some other format you want to convert to a Winkel Tripel projection, the only thing I can think of is using a Winkel Tripel projected Latitude/Longitude image as a guide to drawing the map.

Though I hesitate to provide an explanation for a rather rudely expressed comment, I think the point is that the gradient along edge that results from antialiasing the tilted rectangle isn't continuous. It stays the same value for a span of several pixels, then jumps, instead of changing continuously by a smaller amount. I'd guess that may be due to the number of subsamples used for antialiasing (assuming that's how the antialiasing is achieved). If there are, say, 6 samples per side, there will be only 36 possible levels of gray, which would result in a non-continuous gradient. On the other hand, 16 samples per side would require 256 subsamples, which would probably lead to poor performance.

You ain't just whistlin' Dixie on that. All the different key labels were bad enough, but what really stuck in my craw (to use another down-home expression) was all the different sized keys: regular, medium wide, extra wide, tall, and on and on. (I do think it was a good theme, despite certain tiresome aspects.) Congratulations to @Pixey and @lynxster4, and @AndrewDavid too!

At the risk of becoming tiresome on this subject (supposing I haven't already), I think that curves joined end to end are practically the default use for the Line/Cure tool, so there should be some option to support it beyond "having to zoom right in to get a perfect joining" (to quote the original comment). I must admit, though, that I haven't gotten much support for the idea from other forum members. On the other hand, in just the last week or so there has been this request, along with a request for more than four control points, which would be at least partly addressed by a convenient method for joining multiple cubic curves.

I agree it would be nice if there were, but unfortunately there isn't.

You might also try Ed Harvey's Color Filter plugin.

Try filling a layer above the image layer with the filter color, then setting the layer's blend mode to Multiply. Use Hue/Saturation in the Adjustment menu to modify the layer's color till you get something you like. The lower the saturation, the less effect on the image color. You can also change the filter layer's opacity. Other blend modes may be worth trying. Merge the layers or flatten the image when done.

No. Each individual Line/Curve is a cubic curve, and cubic curves are determined by exactly four nodes. I've (twice) requested that there be a way to automatically use the end point of the previous curve as the starting point for the next curve, so curves could be easily strung together. I think that would help with what you want to do. There's always the ShapeMaker plugin, which allows for cubic splines with any number of nodes.

Without knowing exactly what you want to do, I can't be certain, but perhaps Paste From Clipboard (with Tiled tiling mode) would help.

The problem is that the fill algorithm which does the Magic Wand selection does a 4-connected, not an 8-connected, fill. That means pixels are only considered to be connected if they touch at the sides, not if they touch at either the sides or the corners. Therefore, no matter what the tolerance, the the pixels at the ends of the selection you show will not be connected -- they only touch at the corners. There are good reasons why the Magic Wand uses a 4-connected fill rather than an 8-connected fill, though obviously that doesn't help you. The only thing that immediately comes to mind is to duplicate the layer, apply a one-pixel blur to the duplicate, use the Magic Wand to select the blurred line, then switch to the non-blurred layer. Then use an Intersect Global Magic Wand selection (with a low tolerance) to select only the black pixels within the selection. Sorry I can't think of something simpler and more reliable.

Unless the objects you're drawing are huge or have hundreds of angles, I'm surprised there's a visually noticeable problem. 0.01 degrees is about 0.0001745 radians, which gives a good idea of the size of the relative error you'd be seeing.

Probably your best hope is pyrochilds's Grid Warp, perhaps along with Liquify.

Something that you might try is using Recolor Using Palette, with the colors you want to use added in consecutive entries in PDN's palette. You could use TR's Custom Palette Matcher instead. It's very similar, but gets the colors from palette files instead of the built-in PDN palette. Both plugins replace each image color with the most similar palette color.

It could be due to the fact that the Magic Wand doesn't make antialiased selections; in other words, pixels are either selected or not selected, so the edge of Magic Wand selections is jaggy. If you look at a magnified view of text, the edges of letters will almost always contains partially colored pixels, which represent pixels that are only partly covered by the letter. With the Magic Wand, these are either fully selected on not selected at all. If that's the problem, the solution is generally to find an alternate method to do what you want to do that doesn't involve the Magic Wand. I don't understand exactly what you're doing well enough to confidently suggest such a method. I hope you can provide a bit more explanation, ideally with example images. However, if at all possible, the text you're pasting should be created in a transparent layer. When text is on a transparent background, you can just select a rectangle that contains the text, copy it to the clipboard, then paste it into a transparent layer above the image layer that you want to add the text to. The text will blend smoothly with the image.