T = WlsToT(wls,[sd])
Produce the identity color matching matrix associated
with the passed wavelength sampling.
Arguments may be either [start delta n] description or
a list of wavelengths.
The second input sd is the standard deviation of a Gaussian
blur function which may be incorporated into the T matrix
This second feature should be used with some
caution, as I haven’t thought it completely through. The
current implementation normalizes each row of T so that the
entries sum to 1. This seems as if it is what we want to
get answers consistent with what we get when we don’t
incorporate wavelength blurring.
7/11/03 dhb A few comments added.