[alma-config]low sidelobes vs correct sampling

[Mark Holdaway]

Frederic,

I think your examples fall into a special case, namely, low density at
the edge of the Fourier plane.  Except for the largest configuration, we
are not so interested in accurately reconstructing the visibilities right
at the edge of the Fourier plane; rather, those visibilities end up
getting weighted down by the restorng beam anyway. The sub-Nyquist
sampling at the edge just ends up helping reduce the sidelobe level.

This group has discussed this case to death, about 1.5 to 2 years ago.
Those far out (u,v) points are basically "sacrificed" to the god of
low sidelobes.

If we need an accurate estimate of those visibilities at the edge of the
Fourier plane coverage, we need to go to the next higher configuration.


	-mark