The tst_spline program was added to the cgcam repository on 10/17/2022.
This routine will create a data sequence that is representative of a
thermal duct. The program prompts the user for the number of points in the
data sequence, which allows for a range of sampling from under-sampled
to over-sampled. The program will then fit the data with a cubic spline
and ask the user to specify an up-sampling factor to be used in writing out
the spline fit function value along with the corresponding values
for the first and second derivatives. The data sequence is written to the
file data.out whereas the spline fit is written to the file
fit.out.
The plots below were generated by specifying 17 points for the data sequence and an up-sampling factor of 6. As shown below, the data sequence is coarsly sampled
but the spline fit to it is nice and smooth.
The first derivative is also smooth
The second derivative is continuous, but is not smooth, since it is effectively rendered with a linear fit.
While a zoomed-in plot of the first derivative may give a non-smooth appearance,
this is just an artifact of the plotting algorithm which simply draws straight line segments between adjacent data points. This effect becomes very hard to see if the sampling interval for the fit is increased by adjusting the up-sampling factor to 20.
The plots below were generated by specifying 17 points for the data sequence and an up-sampling factor of 6. As shown below, the data sequence is coarsly sampled
but the spline fit to it is nice and smooth.
The first derivative is also smooth
The second derivative is continuous, but is not smooth, since it is effectively rendered with a linear fit.
While a zoomed-in plot of the first derivative may give a non-smooth appearance,
this is just an artifact of the plotting algorithm which simply draws straight line segments between adjacent data points. This effect becomes very hard to see if the sampling interval for the fit is increased by adjusting the up-sampling factor to 20.