Re: Interpolation w/ cubic convolution kernel - boundary treatment?

"Clay" <physics@xxxxxxxxxxxxx> wrote in message : The common way with
natural cubic splines is the requirement that the
: 2nd derivative == 0 at the knots.

You mean the knots at the end points.

But basically you have correctly
: noticed that something must give at the endpoints.

Yes, unless you are clever. You can insert a knot or point to make the
first and second derivatives what ever you want at the end points. The down
side is that you need to insert points.

You need more
: constraints than are given.

Yes, the extra points are the constraints. Where to place them is the trick.

Peter Nachtwey

.