CompuServe Thread

#Texture Mapping

14 messages in this thread
#123657From: Chris DonerSep 13, 1994 10:28 PM
This may be a dumb question, but, can I take a photograph of wood, marble or whatever and use this for a texture map? If not how is it done?
#123670From: Gus GrubbaSep 14, 1994 12:24 AM
The only dumb question is the unasked one. You got at least three options: – Get a scanner. With a color scanner you can digitize photos and convert them to any type of image file format and later use as materials. – When developing your film, ask for it to be transferred to PHOTO/CD. You can do that as well with already developed negatives and slides as well. If you have a Photo/CD able CDROM, you can load the images using several different photo/paint tools, resize to whatever you want and convert it to a suitable image file to use as texture maps. – Use the prescanned photos in the CD that is included with 3D Studio. You will find dozens of types of woods, marbles, the works. If that isn't enough, there are several CDROM's on the market just for that. Canned photos to be used as materials and all kinds of different applications.
#123695From: Jonas Ruikis [ADESK]Sep 14, 1994 8:15 AM
Hi Chris.. << ….whatever and use this for a texture map? >> Gus has given you many options for aquiring the texture map source. In the material editor note the texture1 and texture 2 dialog entries. For example, load the material APE. Left-Click on the work APE.CEL, hold the mouse key down, and drag it to the VIEW_IMAGE button. This should start you off. Also, note chapter 12 in the tutorials for a complete overview of the Material Editor. jonas[adesk]
#123740From: Chris DonerSep 14, 1994 11:32 AM
Thanks for the info. Before getting my hands dirty I am trying to get a broad overview of how things work. Here is my current understanding: there are three ways to make a material (this is that same as a texture?) 1. A photograph (is this most common?). 2. A mathmatical description of a surface and 3. just design yourself a bitmap. 2 has the advantage of no lose of resolution as you zoom in and if you wrap it around a curve it will not be distorted in any way it should not be. I guess 1 has disadvantages from size relitive to 2 and 3. Any other considerations? To make a building, for example, should I just take a photograph of a building and wrap it around a wireframe? What do most people do?
#123751From: Jonas Ruikis [ADESK]Sep 14, 1994 1:04 PM
Hi Chris, << To make a building, for example, should I just take a photograph of a building and wrap it around a wireframe? What do most people do? >> That's a loaded question… 🙂 I'd suggest using 3ds to load as many of the examples found on the World Creating Toolkit, render a frame, note the quality and then analyze the materials used in the scene from the materials editor. You'll find many different techniques used. jonas[adesk]
#123793From: david W. mennenohSep 14, 1994 6:08 PM
Chris, You can consider a material to be an object consisting of textures. Kind of. Materials are not seperate from textures as you seem to be thinking. They are really one in the same. You create a material in the editor by combining color, transparency, bitmap pictures and other properties. The pictures can be anything from photos to line art to screengrabs. You then apply the materials you create to the objects in your scence. And, yes you _will_lose resolution as you zoom into a material you're using a bitmap in it. You can prove this to yourself by just zooming into a picture in photoshop or something else. The same is true of 3DS. Taking a picture of a building and mapping it onto a box is probably not the best way to make a building. It might work good for far off objects though or as background elements. If you are going to do a walk through or something you will most likely want to build the model by hand. It will lookd much nicer. Reading the tutorials and doing all of them is the best way to learn the program. Although Video Post is not covered nearly good enough to be very useful. <Sorry Jack> Good luck. Dave Mennenoh
#123855From: Chris DonerSep 15, 1994 12:41 AM
Thanks for the info. But, if you do it by hand how can that produce a better result than a bitmap if the resolution of the bitmap is good enough. And when you say do it by hand what is that, don't you just create a bitmap, or, do you create the description of a bitmap? what is video post?
#123923From: Jonas Ruikis [ADESK]Sep 15, 1994 10:36 AM
Hi ya Chris.. << bitmaps vs. geometry.. >> Let's say you were creating an airplane wing. All of the rivets, overlapping metal and such could be represented as a bitmap and used as a texture within a material definition. On the other hand, you could model the rivets, metal and such as geometry. Rendering times would increase but you could zoom in real close… << what is video post? >> Please…. see and do tutorial 22 (as well as the rest) :-)… jonas[adesk]
#124081From: david W. mennenohSep 16, 1994 6:45 AM
<Pics vs. Models> Chris, Think of it like this, 3DS is resolution independent, you can render images to basically a limitless resolution. (10K) You could make billboards out of these images and they would look perfect blown up that large. This is because 3DS is rendering the geometry mathematically. It's not just resizing a picture and saving it for you. For instance, take your building – If you map your image onto a box and light it, put in a background, put the box on a green landscape, it will most likely look perfect from a distance. You are right in _this_ assumption. A pic mapped onto a box will look very good in this instance. But, what if you want the camera to fly up to a bee, in a flower, in a pot, on the sill of one of the windows in the building. Now you can see why geometry would work better. You could never accomplish that by zooming into a picture. Sooner or later you will just get big pixels. Using 3DS you just get closer and closer to the _object_. Well, good luck. Dave Mennenoh
#124207From: Chris DonerSep 16, 1994 9:25 PM
Is there a disadvange to using bitmaps other than losing resolution when zooming in? What if the viewing angle is different than the camera angle of the bitmap used for the creating the texture. Will the distortion of the bitmap be obvious?
#124363From: Yost GroupSep 18, 1994 10:30 AM
Think of a "bitmap" as a mosaic of tiles applied, say, to the side of a cube. If the camera's far enough away, the picture represented by the mosaic looks fine. If you zoom in, you begin to see the grout. If you look at the mosaic side of the cube from a steep angle, the picture is distorted (much like looking at a large mosaic on a floor). The 3ds Renderer has methods of blurring the edges of the pixels (the tiles) so you don't notice the "grout" if you zoom in too far. The distortion, however, is exactly as it would be if you were looking at a mosaic from a steep angle. – Jack
#124384From: Chris DonerSep 18, 1994 3:14 PM
I understand the zooming issue, but, the turning (for lack of a better word) one is different. If I map the face of a (3D wireframe) building with a bitmap for detail, and if the photograph used for the bitmap was taken straight on, then if I position a camera such that the bitmap is turned relitive to the angle of the starting photograph how much distortion is there? Is this a technique that people use and don't turn the bitmap, or just accept the distortion, or something else? Regarding the "grout" and the blurring, I assume what you are talking about here is a form of blending based on an average of the edge pixals of the two "tiles"? Aside from zooming in and turning, are there other disadvantages of using a photograph for mapping onto surfaces to create detail? Thanks for your help.
#124414From: Yost GroupSep 18, 1994 8:14 PM
The main thing to remember about any bitmap, including those digitized from photographs, is that the bitmap image is two-dimensional, whereas the geometry in 3D Studio is three-dimensional. Thus, if you place a photograph of the face of a building on one side of a box in 3ds, it's just like pasting a photo on a real block. If you look at the block from an angle, you're seeing the photo from an angle, but the dimensions represented in the photo won't change dimensionally. (For example, if there's an inset doorway, the inner walls leading to the door aren't going to change properly with the angle.) In addition, lights in 3D Studio aren't going to cast shadows in the photo details. You'd have to set up your lights to matching angles with the light that was cast in the photo. Usually, texture maps are used to provide patterns in small objects in the scene (such as a wood grain texture map on the objects representing the door molding). If you wanted to use a photo of a building mapped to a block, it might work for small buildings off in the distance. You could even make it work on a large building if the scene is not to be animated, but once you mapped the photo on the building, you'd be locking your camera angle to match the photo — which pretty much limits one of the whole reasons to model a scene in 3d to begin with. – J
#124433From: david W. mennenohSep 18, 1994 10:36 PM
Chris, Jack's explanation of why to use models and not bitmaps was great. If I were you I'd do all the tutorials and you'll see why for yourself. If you know of a local dealer they'd probably be willing to show you also. Good luck – 3D Studio is a fantastic package. Dave Mennenoh