From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>

Date: Thu, 10 May 2007 12:33:41 +0200

To: www-svg@w3.org, bimal.thomas@cgi.com

Message-Id: <200705101233.41872.Dr.O.Hoffmann@gmx.de>

Date: Thu, 10 May 2007 12:33:41 +0200

To: www-svg@w3.org, bimal.thomas@cgi.com

Message-Id: <200705101233.41872.Dr.O.Hoffmann@gmx.de>

Hello, the best you can get is a cubic bezier curve in SVG, this requires control points. The control points are related to the derivative of the curve. To get a 'smooth' curve, the derivatives of the curve fragments have to fit together, this means, the curve has to be continuously differentiable. In general for an arbitrary set of points you need a cubic bezier curve to get a continuously differentiable curve. If you know the derivative at your 20 points, it is easy to determine the control points with minor mathematics, because the parametrization of a cubic bezier curve is well known and therefore the derivative from the parametrization parameter too. If you don't know the derivate, the best you can do is (to believe and) to calculate a spline interpolation. Because it is possible to calculate the control points of a cubic bezier curve from a cubic spline interpolation, the problem is reduced on the question, how to determine the best spline interpolation for your data set. This is a well known mathematical problem. A set of linear equations have to be solved to fit your data to a cubic spline. This linear equation system can be reduced to a problem of matrix diagonalisation. If you have the mathematics for this or a library to do this, you have the best fitting cubic spline for your data set and it is simple mathematics to determine the control points of the cubic beziers. Hope this helps...Received on Thursday, 10 May 2007 10:32:02 UTC

*
This archive was generated by hypermail 2.3.1
: Wednesday, 8 March 2017 09:47:10 UTC
*