Newsgroups: comp.graphics
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!noc.near.net!uunet!mcsun!sunic!aun.uninett.no!trane.uninett.no!ugle.unit.no!lise.unit.no!iharkest
From: iharkest@Lise.Unit.NO (Inge Harkestad)
Subject: Tangent vectors of Kochanek-Bartels splines
Message-ID: <1993Apr28.155100.1435@ugle.unit.no>
Sender: news@ugle.unit.no (NetNews Administrator)
Organization: University of Trondheim
Date: Wed, 28 Apr 93 15:51:00 GMT
Lines: 33

I'm working on a system which uses a given set of 3D key frame
positions (x,y,z) to control an imaginary camera movement.  I'm
using Kochanek-Bartels splines (as described in the SIGGRAPH '84
proceedings) to create a variable number of inbetweens between
the key frames.  I want the inbetweens to be given in the form
(x,y,z,dx,dy,dz) where the last three argumentsa are the x, y and
z component of the viewing direction vector of the camera when
positioned at (x,y,z).

The method presented by Kochanek and Bartels only deals with the
positions of the inbetween view points to be generated.  I've
tried to set the viewing direction at a view point equal to the
chord between the two adjacent view points (which in general are
not key frames), but this causes a sligt discontinuity of the
viewing direction vector at the key frame positions (although
the spatial movement seems to work fine; and I'm quite certain
- I think :) - that I've not simply made an implementation
error...)

Now I wonder if anyone out there has used this spline form for
similar purposes and how they decided the viewing vectors.
I'd appreciate replies to be emailed to me at
  iharkest@lise.unit.no
Anyone else interested in the answer will be sent a summary of
the replies if they contact me.

-- 
 _________     __________________
\\        \\  \\
 \\ N G E  \\==\\ A R K E S T A D     iharkest@lise.unit.no         Comp. Sc.
  \\________\\  \\_______________     NTH (Norwegian Institute of Technology)

          "Some people play hard to get, I play hard to want" (Ford Fairlane)
