Osculating/Hermite Interpolation on sampled data

>you might refer to Olli Niemitalo's paper on polynomial interpolation at
>
>http://www.biochem.oulu.fi/~oniemita/dsp/deip.pdf

Well, yes I've read that one too. But I'm actually just referring to the
general case of regular cubic splines. Cubic splines are second order
osculating (both position, first- and second derivatives are matched at
the knots that connect the piecewise polynomials). It is in this case
quite clear to say that, the interpolation is C2-continuous. If we do not
express constraints on the second derivative when creating the spline (for
example for cubic Hermites), then the interpolation is C1-continuous.

But which degree of continuity is reached using when using the sinc
interpolator? Ideally it is fully continous I guess, since the sinc has
the properties you described. But in practice? Doesn't it depend on
whether it is truncated or windowed? And at which zero crossing this is
done?

Just because the interpolation kernel (or impulse response, in this case
the sinc) is fully continous, does that imply that the resulting
interpolation is fully continous too?

This is confusing. Thanks.
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com