CompuServe Thread

#Real 3D vs. AM

2 messages in this thread
#167658From: Mike KelleyMay 1, 1995 3:50 PM
Forgive me, Bill, but a few other questions came to mind when I was reviewing some of your past threads about modelers. (The following quotes are from a thread you created in December): >I think "only played with it for 15 minutes" is key here. I've used it [Animation >Master], as well as most EVERY modeling system on EVERY platform for quite a >while. And I'm just trying to be helpful in pointing out (from experience) that AM >nor a B-Patch Modeler is going to change your life. There are 2 issues here: A >GOOD patch modeler is just another welcome tool **IN** a complete modeling >environment, but it's limited in usefulness, just as AM is. Fun to play with if you've >never used that type of modeling, difficult to use for flexible & complex object >construction. In looking for my "holy grail" of modeling, Animation Master interested me because of the many favorable comments made by some people here. I was held back because apparently the earlier releases didn't export DXF well, but I guess they've solved that problem. Are you saying that you don't think that modeler is as good as Real 3D for character construction? >Unfortunately, until you are able to better visualize an object's contour lines >(related to "slices"), it's going to be awfully tough to model humanoid figures in >ANY modeling system. With a B-Patch modeler (only), just a detailed human head >would likely require DOZENS of tiny patches, concatenated together & with some >real problems arising with blending at the perimeters. Now imagine…..does that >sound EASY (or EASIER ?) ? The most efficient approach is to define the >contours of the entire object, and then tweak the CV's for fine detail. This is the >method used in the best modelers on the planet, working with isoparms (contour >lines). Right at the moment I'm having great success using Deform/Fit by visualizing the side and front views (and the cross sections). Is this what you mean by object contour? The only thing I DON'T like about D/F is that it's hard to make changes in an object without going back and redoing the views, relofting, etc. Plus the cross section positioning (and limitation of same number of vertices) is not nearly as easy as it could be. And usually I'll get very close to the object I want, but not quite. So what can I do? If I start pulling vertices and/or faces, I lose the smoothing effect of the lofted object. I have all of the Yost routines (such as MShape, etc.) and can do some deformation type manuevers, but then I get an abrupt boundary condition that requires a lot of work to smooth. In short, what I'd really like is a D/F sort of dealy that allows me to see the object as I create the outlines. Can you bottom line this for me? Will Real 3D "change my life?" Would you suspect, on the basis of knowing I can D/F with relative ease, that I will be happy with R3D?
#167748From: M. G. BATCHELORMay 1, 1995 11:02 PM
<< Are you saying that you don't think that modeler is as good as Real 3D for character construction? >> There is no comparision really. One (AM) is like constructing a wall brick-by-brick, and the other is like marking off the footprint, and erecting pre-fab sections. Well, that's a lousy example <g>, but I think you get my point. Also, the quality of the spline surface and exported geometry in AM does not compare favorably. One (AM) has a very few tools, and the other is a toolbox full. << Right at the moment I'm having great success using Deform/Fit by visualizing the side and front views (and the cross sections). Is this what you mean by object contour? The only thing I DON'T like about D/F is that it's hard to make changes in an object without going back and redoing the views, relofting, etc. Plus the cross section positioning (and limitation of same number of vertices) is not nearly as easy as it could be. >> Actually, Deform/Fit is a great tool. But there are some limitations. The 2 profile curves are indeed contour lines as seen from an orthogonal point of view, but must be closed loops. There are also other limitations or obstacles using Deform/Fit. But, the YG did a super job in designing these polygonal tools, which DO allow you to work with splines at some level, although triangles are the result. With a freeform modeler, there are fewer restrictions. I really don't think it's right to take up this forum's bandwidth more on this subject, because I explain these things AT LENGTH in the doc's I uploaded. BTW, a detailed head is best accomplished in 3DS by constructing a number of profiles as seen perpendicular to the eyepoint, then rotating the subject (or eyepoint) a few degrees, and construct the implied contours from that perspective. Then use a SOR path, and place the CS's where they belong around the circular/elliptical path. Most people just think of using only one profile with a SOR. With good "vision" in identifying the contour lines and good construction, you can achieve an almost perfect head right out of the lofter, with minimum vertex-level tweaking necessary in the Editor. << Can you bottom line this for me? Will Real 3D "change my life?" Would you suspect, on the basis of knowing I can D/F with relative ease, that I will be happy with R3D? >> Yes, it will change your life…..it'll make it even more complicated because it's new and a different animal <g>. That's why I'm writing the book for them. But if you like the concept of Deform Fit, then you're a candidate to *begin learning* freeform modeling with splines. In short, a lot of your problems mentioned such as vertex level editing polygonal meshes do lead towards justifying freeform spline modeling…..along with a host of other reasons. But you need to be even more adept at visualizing, planning and executing a model's construction using a freeform spline system, than with a more restricted polygonal system. For example, using NURBS vs a uniform B-Spline system, with all else equal, doesn't mean a thing unless you *know how* to use the additional control over the surface nuances & quality….and that takes learning & practice a step beyond uniform B-Spline surface construction. Regards, BILL