MJW

That has to be the most strained analogy I've ever heard.

The switch icon turns off UBB formatting, showing the raw text with the tags instead. For example, instead of showing bold text, it will show the text preceded by "B" in square brackets, and followed by "/B" in square brackets. Type in some text, apply some formatting, then turn off the switch and you'll see what I mean.

Eli, I believe the original comment is about this forum, not PDN.

I see it as a bug, or at a minimum very unexpected behavior. No one would deny that white transparent and black transparent pixels are different internally, but it by no means follows that the Magic Wand should treat visually-identical pixels differently just because they're internally different. And if they are treated differently, they should be treated as though they're a lot different, since three of the four components are as far apart as they could possibly be. So even a high tolerance should differentiate between them.

I don't know why you can't cut-and-paste from EditpadLite. I cut-and-paste text into the PDN Forum from Notepad and other sources all the time.

What skyoxZ is saying is that it probably shouldn't matter to the Magic Wand whether it's transparent white or transparent black, since they're both full transparency, but it does; with a low tolerance setting, the transparent white pixels are selected while the transparent black pixels aren't. I verified that that's true. It seems to occur with tolerances at or below 4%. (This thread most likely belongs in "Troubleshooting & Bug Reports.")

I once started a thread requesting the option of disabling the selection indications, such as the ant trail, but got no replies to the topic. I can't figure out why. Trying to match at a selection boundary with the selection indication is nearly impossible. I know Paint Shop Pro allows that, and I think Photoshop does also. The only solution I know of is to rely on layers and transparency rather than selections. For example, instead of feathering a selection, copy the selection into another layer, disable the selection, and use Feather Object.

This is a very good plugin to use, because it appears to use the correct algorithm for maximizing transparency while making the new color blend to produce the original color when alpha-blended to erased color. Not every plugin that does something similar does it as well. (Must have been a tough crowd, since the plugin's author got no Reputation points for this useful plugin.)

I assume you want to also change the background, since otherwise it would be white on white, which would be easy to do, but not too useful. Making the background transparent seems the most useful, so I'll assume that's what you want to do. There's a very useful and easy-to-use plugin to do that called Make Transparent. After you've installed the plugin, load the image, set the Primary Color to white, then run Effects>Color>Make Transparent. All the white will be replaced by transparency. (You could also leave the Primary Color as black, run Make Transparent, and use its color-wheel control to set the color to erase to white).) Now to make the font white. Run Adjustments>Hue/Saturation. Move the Lightness control to 100, and the text will be white. (There are other ways to do it, but this one is easy.) EDIT: Here's an easier way. Run Adjustments>Invert Colors to make the text white on a black background. If you then want the background to be transparent, run Make Transparent with the Primary Color set to (the default) black.

Do you just want to put a colored border around the image, or do you have something more elaborate in mind, such as an imitation of a gold frame?

Congratulations to Pixey for her elegant winning entry! I was impressed by the quality of the other entries, and pleased by the wide variety of approaches used for the theme. Many thanks to Drew for for hosting the competition.

I don't really agree that HSV Eraser is less accurate. If an erased version is set against a layer with the original background color, the result is in almost every case visually indistinguishable from the original. The alpha-blended version may be up to one bit different in some color components due to rounding, but is otherwise identical. You could achieve "more accurate" results by only removing pixels that exactly match the background color, while leaving the remaining pixels unaltered and opaque, but that would be contrary to the objective of removing the background color as if it were a layer that was blended against.The goal of HSV Eraser's algorithm is to make the colors as transparent as possible. EDIT: I may possibly be able to reduce the number of one-bit errors by slightly modifying the computation. I'll probably look into that. EDIT 2: As mentioned in an edit to my original comment, I really messed up my original attempt, and have corrected the mistakes. I now believe Make Transparent probably uses the same method as HSV Eraser to remove as much of the background color as possible.

Perhaps you could try the HSV Eraser. Set the color to erase to the red background color. The easiest was to do that is to use the Color Picker tool to set the Primary Color, then choose Primary Color as the HSV Eraser Match Color. Set the Hue Tolerance to 1.0, the Portion of Non-Erased Color to Preserve to 1.0, and check Gray Matches All Hues. . Example (shown against PDN's checkerboard background): EDIT: Not to be immodest, but HSV Eraser's algorithm is better than Grim Reaper's; it removes the background color more completely. I'm not sure about Make Transparent, but my erased version seems to have more of the background color eliminated. When Portion of Non-Erased Color to Preserve is set to 1.0, HSV Eraser makes the colors as transparent as possible, while still producing the original colors when alpha blended to the original background color. EDIT 2: I realized I had forgotten I needed to set the Hue Tolerance to 1.0 and to check Gray Matches All Hues, which meant that some of the pixels were only partially erased. I fixed the instructions, and substituted an updated image. After comparing the results now that I'm doing it correctly, I believe Make Transparent uses the same basic algorithm as HSV Eraser. Since Make Transparent is considerably easier to use for this specific task, I recommended using it.

Another possibility is TR's Paste Warp. As a side note, there are two common transformations for mapping rectangles into trapezoids: perspective and bilinear. Each has advantages and disadvantages. Perspective transformations map lines into lines, which is very desirable for tapering text; however, they also compress the text in both directions, so it doesn't look tapered so much as (naturally enough) viewed at an angle. Bilinear transformations can compress the text in only one direction, but they have the decided disadvantage that diagonal straight lines are transformed into curves.

Looks like Arial to me, with the lower line in bold face.

My second entry. As is obvious, the image was derived from Théophile Steinlen's famous "Le Chat Noir" poster I use as my avatar.

Alpha is transparency (or perhaps you could say opaqueness). The higher the alpha value, the more opaque the pixel. A 0 alpha pixel is completely transparent; a 255 alpha pixel is opaque. The default mode for Paste Alpha is to paste the grayness of the pixel in the clipboard as the alpha in the image. If the clipboard pixel is black, the corresponding image pixel will be made transparent; if it's white, the image pixel will be made opaque. This can be reversed by checking the Invert Calculation checkbox. You should make a gradient image which is black where you want one image to show, white where you want the other image to show, and shades of gray in the transition region. Copy the gradient to the clipboard, then (with both image layers visible) select the top layer and run Paste Alpha. The transparency should either be what you want, or reversed from what you want. If it's reversed, check Invert Calculation. (Note, I could tell you which layer to put on top, but it's easier and less confusing to just put one of the layers on top, run Paste Alpha on that layer, and if the transparency is opposite from what it should be, check Invert Calculation.) Click the question mark in the upper-right frame of Paste Alpha to get a concise Help Menu.

What is it about Paste Alpha that's causing problems? As long as the gradient layer is the same size as the image layer, it should just be a matter of copying the gradient into the clipboard, selecting the topmost image layer, then running Paste Alpha. The Invert Calculation checkbox can be used if the alpha is opposite of what you want, so it doesn't even matter whether the gradient is white-on-black or black-on-white, or which of the two image layers is on top.

I suggest making a black and white version of the gradient you want, then using BoltBait's Paste Alpha plugin to transfer the gray tones into the top layer as alphas.

This is an excellent plugin, but there is one additional feature I wish it had: the option of using the image in the clipboard in place of a file. That would make editing the fill images easier. EDIT: While I'm at it (making a pest of myself), I'll mention two other other changes I'd like. First, I wish the rotation range were always a positive value that represented a +/- range., so that a 0 Rotation would produce objects that on the average point in the original direction no matter what the range. Second, I wish the Shrink Max were a value from some small value to 1, with that value being the maximum reduction from the scaled size. The default would be 1. If the value were 0.5, the smallest fill objects would be half the scaled size. That seems to me to be more intuitive. (Perhaps it would need to be renamed.)

In the water-drop tutorials using Random Shape Fill, it's interesting to substitute TR's Custom Random Filler with some drip-like shapes in place of the circles. I suggest some round shapes along with some more teardrop-shaped drips, all against a transparent background. Make sure the rotation keeps the drops dripping downward, but with a bit of variation, and that there's some range to the shrinkage for variety. Round shapes are probably best for the smaller drops, and the teardrop shapes for the larger drops, since gravity overcomes surface tension as the size increases.

I may be able to offer some useful suggestions, but I need to know more what you want to achieve. For example (and perhaps primarily), do you want to have an identifiable image behind the watery texture, or do you just want the general impression of light and dark areas?

I've never used Layer>Import From File (I likely will, though, now that I know about it), but this sequence has a small misconception. There's no need to add a new layer, because Import From File creates a new layer for the image it imports; in fact, if you import a multi-layer .pdn file, it will add a layer for each layer in the imported image. If you first add a layer, that layer will be ignored by the importation, so it will be empty.

I just tried the plugin with a quite large (2.35 MB) SVG file output by Inkscape, and it worked great. Now that I understand better how to convert raster images to vector images with Inkscape, this may prove to be quite useful.

I'm probably once again repeating what's already been said (I'm too lazy to read all the other responses in detail), but perhaps your problem is how you're combining the two images. Looking at the second example of the shiny button, there's an outer metallic ring, and an inner glass-like button. The edge around the inner button is very jagged. Perhaps you combined the images by selecting the button from its layer then pasting it into the metallic ring layer. If so, that's not, in my opinion, how you should do it. Images should be combined by putting them in separate (transparent) layers, then merging the layers. If you want to modify the size of an element, use Paste into New Layer. Adjust the size of the selection, then after doing any antialiasing steps, merge down. Having the element -- e.g., the glass button -- in its own layer, surrounded by transparency, allows you to use AA's Assistant, and BoltBait's Effects>Object>Feather Object to soften the edges before you merge. Both those smoothing plugins somewhat reduce the size of the object they're applied to, so make sure there's something underneath. For instance, make the metallic ring extend inward far enough that it will fill in the small region that results from slightly reducing the button size. (If the button image wasn't originally produced by cutting and pasting the inner button, but was instead produced by creating the metal ring and glass button on separate transparent layers, then merging, just follow that same procedure, but before merging, smooth the inner button edge using AA's Assistant and Feather Object.)