Attached Link: http://home.pb.net/~stevecox/boxtut.htm

Michalki - Something to remember is that a photograph is itself simply a planar projection. Even if you flatten out the facets of the model on the texture map, a single photograph would not be able to take advantage of that. To get the best results, photograph the source from the same direction that you plan to use when rendering the model. Then do a planar projection on the model from the same angle, apply the photograph and render. To get better results from several angles, you can take photographs from 6 directions (front, back, left, right, top and bottom, or 4 sides if you don't care about top/bottom) use box mapping on the model, drop the photographs onto the texture map template and render (the attached link is for a tutorial where I use this technique to model a box, admittedly a best case scenario =) Lastly, you can use the "select by vertex" feature of UVMapper to select and move individual vertices to get the layout exactly how you want it. While this can be tedious, and you'll need to fill in the "spread" facets by hand, the results can be spectacular. Hope this helps, Steve