CompuServe Messages

More on mapping coordinates: Imagine someone mails you a picture of a logo, a model of, say, a chair, and a letter asking you to paint the logo onto the chair model. That's all the letter says. But you don't know how big the logo should be, or where on the chair it should be, or in what orientation, or whether it should be painted only once or in a repeating pattern. 3DS has the same problem if you just give it a shape and a material with maps. It has no way of knowing HOW to apply the maps to the object. This is what mapping coordinates are for. As I said before, there are seven types: the three manual types (Rectangular, Cylinderical, and Spherical), the more-automated Lofted, Box and Face mappings, plus IPAS .PXP/.AXP-modified mapping. Starting with the manual types: Rectangular mapping works as if one took an insubstantial rectangular picture that was magically enchanted so that anything it passed through would take on the colors of the picture, and you then passed the rectangle over an object. Each pixel of the picture that intersects the surface of the object causes that tiny spot of the object to take on that color (or opacity or bump deformation or whatever). Another way of looking at it is to imagine a rectangular picture wherein each pixel casts a ray of light in both directions from it, exactly perpendicular to the rectangle itself, and extending to infinity in both directions. Where any of these rays intercept the mapped object(s) or element(s), they paint the surface with their individual pixel colors. Cylinderical mapping imagines a picture rolled up into a cylinder. This time, think of it as rays of light being emitted from the surface of the cylinder both inwards towards the center and outwards towards infinity, but always perpendicular to a plane tangent to the cylinder at the point the pixel in question. Again, any object so mapped which intersects these rays gets painted on accordingly. Usually, a cylinderical map is positioned so that its center and axis match up with the center and an axis of the object(s) or element(s) to be mapped, but this need not always be the case. Interesting mapping effects can be done by having the cylinder off-center or even completely outside the object. For spherical mapping, imagine the picture wrapped around a globe and projecting its rays of colorization inwards towards the center and outwards to infinity, with each ray being precisely perpendicular to a plane tangent to the sphere at the pixel. Again, the center of mapping need not, but usually does, coincide with the center of the mapped object. You can "tile" any of the manual types, causing the picture to repeat horizontally and/or vertically. Note that, for rectangular, the aspect ratio of the mapping icon should be as close as possible to the aspect ration of the picture it contains, or else the picture will appear distorted (compressed or expanded). Lofted mapping basically follows the loft path in the 3D Lofter and applies coordinates to the resulting object so that the pictures appear wrapped perfectly around the object. You can set tiling options here as well. Box mapping is like rectangular mapping but works semi-automatically. You simply specify a material for the Front, Back, Left, Right, Top, and Bottom of the object (as seen from its local coordinates — to verify those, do an Object Pivot on it from the Keyframer), and those are applied with mapping coordinates that act as if the rectangles for each of the three dimensions were exactly sized to precisely enclose the object, but no larger than that. Face mapping applies mapping coordinates to each individual face or group of coplanar faces. This is set on the material layer and overrides all other mapping coordinates. It results in mapped patterns that completely cover the surface, mapped to each face. The results aren't easily predictable but can be quite beautiful. Check out the sample FACEMAP.3DS file, and apply the sample FACEMAP METAL BUMP material to the polyhedron. Finally, certain .PXPs and .AXPs can apply their own more sophisticated mapping at modeling (.PXP) or rendering (.AXP) time. In the case of .AXPs, you must first apply any arbitrary mapping coordinates (default rectangular will work fine) to the .AXP stand-in object so that the Renderer doesn't abort with an "Object needs mapping coordinates" error before the .AXP even gets a chance to run. Those arbitrary coordinates will usually be completely replaced by the IPAS-generated coordinates, though, but it may use them as a basis. For example, DISINTEGRATE chooses the color of each particle by the dominant color of the original mapping of the object at that point. FIREWORKS, on the other hand, completely ignores the original coordinates and instead uses the applied map as a time/distance color chart to control the color of each of the fireworks particles and their tails over time.