MJW

Okay, now I understand better. The short answer is that that's not what the Texture Merger or Texture Shader do. You need to have a Height Map. For the egg, preferably a 24-bit map, which means that the R,G, and B channels are treated as a single number. The Texture Merger is intended to make it easier to produce height maps. For an egg, I'd start with a sphere height map, like this: You can use the Texture Merger to make it oval instead of round. First, copy it to the clipboard Run the Texture Merger. Set the Merge Method to Clipboard (No Displacement). Set the XY Proportion to 0.12. Set the Merged Height Scale to 240. (Note: this only affects the shading, not the final height map.) Enable Show Shading, and you will be able to see what the shaded height map looks like. Now you can disable Show Shading and exit the Texture Merger and you'll have an egg height map. If you want to shade the egg green, set the Primary Color to green and run the Texture Shader. Set the Image to Primary Color. (Image is the image that's mapped onto the height map. It can be a fixed color, or from the clipboard.) Set the Texture Height to 240. Position the Directional Light to improve the results. (The default straight-on light direction is pretty boring!) You can adjust the speculariy controls if you want to make the egg shinier. If you want to make an Easter egg, you can put an pattern in the clipboard and set the Image to Clipboard.

I think it would probably help to start with the Texture Shader plugin. In terms of complexity, the Texture Merger is like the Texture Shader squared. Nevertheless, I'll briefly explain mapping displacement and merged height maps. A merged height map is a combination of two different height maps. There are a number of ways they can be combined. The most common is probably to add a texture, like a leather texture, to an object, such as a sphere. The goal is to produce a sphere that appears to be covered with leather, like a basketball. The Texture Merger is intended to allow textures to be combined. Mapping Displacement is the way the object affects the image that's mapped onto it. I'll use Texture Shader examples. If you paint an image onto a sphere, the image bulges, which is what makes the sphere look round. That's what the Surface Offset mapping modes try to do. On the other hand, if a sphere reflects an image in front of it, the result will be very different. That's what the various Reflection mapping modes try to do. In each case, the image in the clipboard is distorted by the height map in the canvas. That distortion is what I call the mapping displacement. Note: The default Mapping Mode for the Texture Merger is None, which is probably why the object didn't look round. The default mode for the Texture Shader is Surface Offset 1, so it will probably look round with the default settings. You also need to make sure the Texture Height is appropriate. If it's too small, the object will look flat.

Yes, that should work.

I'd be happy to do so. For the watch face: I extensively used xod's Clock plugin, with lots of by-hand modifications, such as eliminating the numbers ending in five on the small dial. I did the tick marks separately from the numbers to get the positioning I liked. The part I like best is the way the outer (red) numbers go upside down at the bottom of the dial. I did that by first using MadJik's Polar Transformation plugin to transform the dial into rectangular coordinates to provide a guide for positioning. I typed the numbers at the proper positions, then used Polar Transformation to transform the numbers to polar coordinates. I drew the hands using the Line/Curve tool. I added a light drop shadow with KrisVDM's Drop Shadow. The circular indentation on the dial was done using the Texture Shader. The main number font (of which I'm quite fond) is actually a blackletter font called DS Walbaumfraktur. I think the little curls in the 2s and 3s give it a nice turn-of-the-century look (20th century, not 13th). For the watch case: This was almost all done using the Texture Shader and Texture Merger plugins. The metal case and the glass lens are essentially flattened spheres, with the Texture Merger's Curvature control used to modify the shape. The watch face was mapped onto the lens with the Texture Shader, using the Refraction mapping mode. The part of the case I like best is the gridded ring. To do that, I made a white-on-black rectangular grid, then blacked out everything but one row (so it was sort of like a railroad track). I then used Polar Transformation to make it circular, slightly blurred it to smooth the transitions, and mapped it onto the case with the Texture Merger using Subtract mode. It took a lot of playing around to get the right proportions! For the winding stem and knob: The winding knob is just a slightly horizontally-stretched sphere with a grid of vertical lines mapped onto it with the Texture Merger. The cylindrical part was a bit tricky. I started with 24-bit sphere height map. I selected a pixel-wide vertical rectangular slice, then used the Move Pixel tool to stretch it horizontally. That worked really well to form a cylinder for the stem and the ring. The ring is the part I like best, mostly because I had no idea how I was going to do it, but it turned out to be fairly easy. I just elongated the (24-bit) cylinder height map, and used Polar Transformation to transform it into polar coordinates. Because Polar Transformation is meant to deal with RGB, not height maps, the result had an uneven-looking surface caused by the faulty-for-24-bit-values interpolation (though less so than I expected). I used the Texture Smoother plugin to smooth it. The bottom of the stem bulges then narrows. The bulge was made by Adding a thin horizontal cylinder with the Texture Merger, and the indentation was done by Subtracting a thin cylinder. The exact details for the indentation are a bit complex; specifically, the way the profile is indented. When the cylinder was initially subtracted, the profile indentations were not transparent, but had a Z depth of 0, due to clamping. I used the Feather feature of the Texture Scaler to replace the 0 depth with transparency. The "ears" the ring goes through are very elongated (and possibly tilted -- I don't remember) spheres. I combined all the elements of the case with the Texture Merger before shading, using the Maximum merge mode. I shaded it with the Texture Shader using the Reflection (with Re-Reflection) shading mode. For the clipboard image (which is what is reflected), I used a sort of swirly pattern I made a long time ago for the Decorative Glass SOTW. I set the color of the lights to yellowish-orange so it looked like gold. The shading had one feature I hadn't expected, but was very pleased about. The winding stem and knob seem to be reflected onto the watch case. This is, I suppose, due to the way the Reflection (with re-Reflection) mode works. I wasn't completely happy with the case's gold coloring, which seemed a little overdone, so I recolored it using my HSV Scrambler plugin. I set the source for all the elements to Value. For the Hue, I set the Offset to produce a golden color, and I set the Scale factor to a quite small value, to add a slight variation, I set the Saturation controls so the saturation would decrease slightly as the Value increased. I adjusted to Value to get the contrast that looked good to me. One useful thing I discovered is that with the correct adjustments, I could get a very nice silver color. I liked it so well, I considered changing it to a silver watch, but stayed with gold. (I just realized all the features I like best -- the outside red numbers, gridded ring design on the case, and winding-stem ring -- involved the Polar Transformation plugin! Thanks, MadJik!)

I don't quite understand how you intend to use the green layer for two different values. What you could do is have R, G, and B layers, and a grey-scale layer for alpha. You could merge the R, G, and B layers using Additive blending, then copy the grey layer to the clipboard, and merge it into the color layer using BoltBait's Paste Alpha plugin to paste the gray as alpha. That's not as convenient as just merging layers, but unfortunately there's no way of merging that will just transfer the alpha value.

I really don't like that idea. I was very disappointed that Woodsy couldn't submit a pocket watch entry.

If it should be limited to one or the other of toe_head2001's examples, I think the combination wrench is a good choice. It provides a bit more variety than a double-ended wrench in the first example. Glad to see you back in the thick of things, Woodsy!

Though I was considering trying an adjustable crescent wrench (thinking the knurled adjuster screw would be an interesting challenge), I think it should be limited to a fixed-sized wrench. That seems more consistent with the original concept of the OOTF competition of everyone trying to do the best version of the same object.

Though it doesn't change the point being made that sometimes the layer order matters, I believe Difference blend mode uses the absolute value of the difference, so for two opaque layers, the order shouldn't matter.

Thanks, Pixey and Red ochre! I hope my comment didn't come off as groveling for praise (or at least that it isn't too obvious). And Pixey, I'm very impressed by your use of a hunter case, and especially the off-angle POV.

Were they drawn on a transparent background? Align Object aligns "objects," where an object is a non-transparent region surrounded by transparency. (Also, as with all other effects, Align Object only affects the currently active layer.)

Congratulations to Pixey! Immodest though I may be, I must say, I really liked my image. I think it's one of the best I've done.

I should have read the other comments, I guess!

Eli, that's impressive, but why are the colors yellow, cyan, and magenta rather than the light primaries, yellow, green, and red?

I and quite a few others are of the opposite opinion: Making a layer visible shouldn't make it the active layer, and making it invisible shouldn't make it inactive. They should be independent. I often make a layer visible to see how it combines with the layer I'm working on, with no intention of changing it.

Congratulations to Pixey and all the other entrants. I'm embarrassed to admit I forgot to vote. I fully intended to, but was debating which of two entries would be my final selection. I decided to decide later, then forgot. I actually remembered this morning, apparently just after the voting closed. I'll try not to let it happen again. Some very nicely done entries. I think dipstick and Beta0 did especially good jobs of capturing the cubist look, in very different ways.

You could try my Color Clearer plugin, which will convert any color to transparency (in the sense that if you blend it to a background of the selected color, the picture will be unchanged).

I don't want to say you're wrong, but the behavior on merging is exactly what would be expected. If you have a partially transparent layer on top of an opaque layer, then before merging the combination will be opaque: none of the lower layers will show through. So the merged layer is also opaque. The alpha-blending equation is the sum of the alphas (considered as values between 0 and 1) minus their product. If either layer is opaque, the merged layer will be opaque (1 + a - 1 * a = 1). EDIT: Modified because I wasn't certain what some of what I'd said was precisely true.

I'd also suggest another of Red ochre's plugins, Gears. To draw circles, set: Tooth shape: None Shaft hole radius: 0 Spoke shape: None Line width: The desired line width Object: Unchecked Use Tooth width or Overall size to set the radius. Tooth width is better, I believe, because it maintains proportionality: doubling the Tooth width doubles the circle's radius. (Concentric circles can, of course, be drawn by repeating the plugin with the same center but different radii.) One disadvantage is that the radius or diameter can't be set directly in pixels. If Overall size is set to 12.1, the radius will very close to ten times the Tooth width. If Overall size is set to -20.75, the diameter will be very close to ten times the Tooth width. Alternately, if the Number of teeth is changed (from the default of 100) to 126, and the Overall size is left at 0, the radius will be close to ten times the Tooth width. if the Number of teeth is changed to 63, and the Overall size is left at 0, the diameter will be close to ten times the Tooth width.

How nearly does it have to match? The font looks a fair amount like italic Palatino Linotype, though Palatino Linotype is in my opinion more elegant looking (and better spaced). The lowercase letters are quite a close match; the uppercase "G" and "D" in your font are somewhat narrower. I suspect that unless your font is very commonly used, it will be almost impossible to find an exact match. There are many, many similar fonts. The lowercase "w" is somewhat distinctive in the way it's rounded and the degree to which the final stroke curls back, so that's one of the first things to compare. It probably goes without saying that it wouldn't hurt to have a better example image than a small, blurry one, with a limited set of letters, including only two uppercase. That's probably why the automatic matches failed. (I put this comment in italic Georgia, which is another similar font: "Greece seems weak. Declares war on Greece.")

We very much appreciate you anyway, toe_head2001! I see the entries are now posted in full size, which I also appreciate. I do think it would probably still be a good idea to reduce the maximum entry size to 600x600 in future competitions. 800x800 is pretty large.

Try toe_head2001's Quadrilateral Correction plugin. EDIT: I didn't see that toe_head2001 had already replied. I guess it was because it was on the next page.

I notice that when the entries are posted for voting, they're posted at a reduced size. If they can't be posted at the 800x800 size allowed by the rules, I think the rules should be changed to the maximum size for the voted-on version. I don't see much point in allowing an entry size larger than the version that will be voted on. (I wouldn't have problem with the maximum size for entries being reduced to perhaps 600x600.)

Use MadJik's Polar Transformation plugin with "transform rectangular to polar" selected. It will be under the Distort menu. It's kind of a finicky plugin: getting the exact result you're after may take a little experimentation with the scaling. I recommend using a square canvas, with the object to be transformed centered in the window. In order to get the half circle you want, make the object's length half the canvas width.

Mejallen, you did an excellent job! "Heavenly" looking, without going over the top. The flare in place of the halo was a good choice. It conveys the same idea, but more subtly.