pyle@lll-lcc.UUCP (04/10/87)
Has anyone scan converted quadric surfaces? Has anyone done it using the methods described in "The Algebraic Properties of Homogeneous Second Order Surfaces", by James Blinn? My initial problem is that the matrix describing the quadric surface in screen space becomes very large (or rather) the numbers within the matrix become very large. When I try to calculate the matrix that represents the silhouette curve of the quadric, the numbers are so big that it overflows the precision of the machine. It could be that I am not calculating my adjoint matrices correctly. Please drop me a note if you can. Thanks, Ernie Pyle
pyle@lll-lcc.UUCP (06/18/87)
I submitted this request once before, but I would like to submit it again. Does anyone know of a general way of scan-convert quadric surfaces? I have had much sucess using Blinn's methods (see "The Algebric Properties of Homogeneous Second Order Surfaces", James F. Blinn), but his methods appear to fail on cones and cylinders (it's hard to determine the sillouette plane). I have read two papers that seem to indicate that uniform processing of quadrics is possible. The papers are: "Visible Surface Algorithms for Quadric Patches", Robert Mahl, IEEE TOC, Jan. 1972. "A Parametric Algorithm for Drawing Pictures of Solid Objects Composed of Quadric Surfaces", Joshua Levin, Comm. of the ACM, Oct. 1976. Any help is appreciated. Thanks, Ernie Pyle