Page 1 of 1

invert's interpolation kernel?

Posted: Fri Nov 13, 2015 8:22 pm
by luke
Hi, does anyone know if there is documentation describing invert's uv gridding process? In particular, describing the choice of gridding interpolation kernel? Also, is the interpolation kernel different from CASA's?

I have looked in the manual, as well as papers describing MIRIAD, and had a quick look in the source code, but nothing states the details of the process. I know that "Interferometry and synthesis in radio astronomy, A. R. Thompson James M Moran; George W Swenson" suggests that prolate spheroidal wavefunctions are the common choice, but even then, does not suggest how many nearest neighbours are used for interpolation.

Thanks,
Luke

Re: invert's interpolation kernel?

Posted: Wed Nov 18, 2015 2:23 am
by luke
Just posting an update for anyone else wanting to know the same information.

The MIRIAD gridding process is coded within mapper.for and grid.for, mapper.for performs the gridding, and grid.for constructs the interpolation kernel. It looks like spheroidal is the choice of interpolation kernel.

The reference given for the gridding process is, "F.R. Schwab "Optimal gridding of visibility data in radio interferometry", Indirect Imaging, J.A. Roberts, ed., Cambridge Univ. Press, 1984."