I've decided to make my first game a 3D Volleyball game. I'm using Milkshape 3D to build 3D objects for the game. The biggest problem I'm having right now is creating a texture for the ball. Since the ball model is a sphere, the texture must be a mercator projection of the ball's surface (like the 2D images of the earth that are used in some demos). I can't quite figure out the best way to create such a texture. In an experiment with a tennis ball I tried spinning it around and taking a bunch of pictures so I could later assemble the centers of the pictures into a crude mercator projection. The problem with this approach is that you have to take lots of pictures, plus the poles of rotation are not quite visible to the camera. Futhermore, the distortion introduced at the poles by using a mercator texture might not be good for a volleyball model (a volleyball is more symmetric than the tennis ball I used in my tests). I've also tried texturing individual groups of faces (dividing the volleyball's surface into six groups of faces), but the stacks/slices sphere that Milkshape uses does not lend itself to texturing a volleyball in this manner. Can anybody offer me advice on how to create a volleyball texture? Would it be any easier to use professional modeling software like Lightwave, Maya or 3D Studio? [edited by - chronos on September 30, 2002 10:22:04 PM]

I just bought a volleyball and took a bunch of pictures of it. Sixty-four pictures comprising a single rotation of the ball (I think I understand Ray Harryhausen a little better now). The next thing I'll have to do is to assemble together the centers of the different images into the actual texture.

Is there an easier way to texture a volleyball? How would you do it? There ought to be an easier way.

[edited by - chronos on October 2, 2002 3:57:17 AM]

Search the board somewhat. I don''t recall the name of the post, but someone found a plugin that would allow photoshop to create a spherical texture for you using a regular image. Sorry I can''t be of more help.

Regards.

Thanks Neostorm. Unfortunately the search feature is currently disabled, but I appreciate your help.

After trying to assemble together a couple of the pictures I took I decided it's too much work and besides would not yield an ideal picture because of the width of the slices used. I'm going to try a different method:

First, use small stickers to identify the two points along the ball that will become the poles in your model. Next, use a measuring tape to place a bunch of stickers every so many units around the ball's equator line. Next, place the measuring tape over one of the lines that goes from one pole to another across one of the points along the equator and measure where the ball's features intersect this line. Do the same for every point along the ball's equator. Next, plot a point for every measurement taken and interpolate the values in between (perhaps by hand); this will be the outline for your texture image.

I never imagined this would be so hard.

[edited by - chronos on October 3, 2002 3:32:06 PM]

It turns out the mercator projection is not what is used to produce the map that is wrapped around a sphere to create a model of the earth. This page discusses spherical projections and uses the phrase "direct polar" to refer to the images used as textures for models of the earth. It seems I had the two mixed up.

I''m still hoping for some advice. Hasn''t anybody here ever created a texture to represent the surface of a ball?

maybe it''s not the help you''re looking for, but my reaction is to model the volleyball in a way that eases texturing. in a sense, a volleyball is like a cube, at least in terms of the patterns on it. try making a cube, subdividing the faces a bit, and then ''sphere-izing'' it. that way you can map it like a box, with a texture that is essentially 3 stripes.

wendigo23,

You're right, a volleyball is almost like a box. I just tried the method you described and it works great. I think I'll continue using the method I described earlier, since I've already plotted some points onto graph paper and it looks very promising. It's certainly very interesting to hear about other methods so as to better judge the value of my own. Thanks for replying.

[edited by - chronos on October 4, 2002 5:36:50 PM]