[daip] Query regarding uv-griding in AIPS

[Eric Greisen]

On 08/22/2017 05:00 AM, Prasun Dutta, Assistant Professor, Dept. of 
Physics, IIT (BHU) wrote:
> Dear Eric,
> 
> I had a few quaries regarding IMAGR in AIPS and after soe research, 
> thought of writing to you directly. Here are they.
> 
> 1. In order to make the (dirty) images, the visibility data is gridded. 
> The section 3.3 in Chapter 7 (Imaging) of the book "Synthesis Imaging in 
> Radio Astronomy II" refers to several gridding algorithms. I was 
> wondering which of this is used in IMAGR by default.  Also is there a 
> way to change the gridding method ( using a verb/parameter) in the IMAGR.
> 
> 2. I was involved in CLEANing an extend source and was using MINPATCH=51 
> as recommended in the help section. I understand what the parameter does 
> and why it is recommended to have a high value. I also read form there 
> the following:
> 

In IMAGR (and UVMAP) the convolving function is chosen with adverbs 
XTYPE, XPARM, YTYPE, and YPARM.  You are free to play with them if you 
like.  The default is the spheroidal function with parameters found by 
Schwab & Greisen a very long time ago.  I suspect no one plays with 
these adverbs - they can do very bad things to your images.  HELP 
UVnTYPE will give you information about type n.

A MINPATCH of 51 is actually fairly small and I suspect it was chosen 
when memory limitations were serious (in our Array Processors).  The 
programs actually use as much as they can given the histogram of the 
beam and the nature of the images.  The greater the accuracy of the 
subtraction in the minor cycles, the greater the accuracy in the 
components chosen.  Note that for extended sources you might want to 
consider the multi-scale Clean (NGAUSS>1, WGAUSS=0,x,y...) available in 
IMAGR.

Eric Greisen