Wednesday, April 11, 2007

Imaging with Multiple Fields

There are two reasons you might want to image with multiple facets.
1) To correct for 3D effects, and image degradation that goes along with that.
2) Because there's a really bright source far away from the center of the field that's messing up your image, but you don't want to image all the way out to it. It would be so much nicer if you could just put a small field on it!

Here's some of the tricks of imaging with multiple fields:

You use SETFC to figure out where to put the facets. I think SETFC is really a pretty great program.
The first time you run it, if you set cellsize and imsize to 0, SETFC will actually recommend a pixel size and image size to you! Like this:

1 5 23-MAR-2007 14:08:03 SETFC Task SETFC (release of 31DEC05) begins
1 3 23-MAR-2007 14:08:03 SETFC Found NGC3184 SPLIT Seq 1 Disk: 1 in slot 6
1 4 23-MAR-2007 14:08:20 SETFC SETCEL: recommends IMSIZE 932 CELLSIZE 1.26446
1 4 23-MAR-2007 14:08:20 SETFC SETCEL: returns IMSIZE 1024 CELLSIZE 1.17417
1 2 23-MAR-2007 14:08:20 SETFC ZTXOP2: using translated file name =
1 2 23-MAR-2007 14:08:20 SETFC ZTXOP2: TEST
1 4 23-MAR-2007 14:08:20 SETFC FLYEYE added 7 fields to BOXFILE to .283 deg
1 3 23-MAR-2007 14:08:20 SETFC Searching catalog between .28 and 1.50 degrees radius
1 2 23-MAR-2007 14:08:20 SETFC ZTXOP2: using translated file name =

I often rerun SETFC with a rounded off version of cellsize (like 1.2 in this case) so I can remember it more easily. CHANGE AS OF 5/09: I"ve really decided that the CELLSIZE suggested by SETFC is far too small. It often just barely Nyquist samples your source-- often giving only 2 or 3 pixels across it. I now try to use CELLSIZEs which are 2 or 3 times smaller than that suggested by SETFC. Also, I don't like imaging over fields bigger than 2048 because than IMAGR averages many pixels together to display the field, and it's hard to tell what's going on. So, I often have imsize of 2048 and a cellsize that gives ~5--6 pixels across a point source.

Now, the other cool thing about SETFC is that you can set an inner tiling of facets, and then you can have some external fields on bright sources (SETFC queries the NVSS database to know where there are bright sources around your source). If you want both internal and outlying fields, you're input will look like this:

AIPS 1: SETFC: Task to make a BOXFILE for input to IMAGR
AIPS 1: Adverbs Values Comments
AIPS 1: ----------------------------------------------------------------
AIPS 1: INNAME 'NGC' UV dataset name (name)
AIPS 1: INCLASS 'SPLIT' UV dataset name (class)
AIPS 1: INSEQ 1 UV dataset name (seq. #)
AIPS 1: INDISK 1 Disk drive #
AIPS 1: SOURCES *all ' ' Source selected
AIPS 1: BCOUNT 1 First field number to use
AIPS 1: disk file to write to (the
AIPS 1: input BOXFILE for IMAGR)
AIPS 1: CELLSIZE 1.2 1.2 (X,Y) size of grid in asec
AIPS 1: IMSIZE 1024 1024 field size
AIPS 1: SHIFT 0 0 Position shift (RA,Dec) asec
AIPS 1: for all fields
AIPS 1: FLUX 0 Minimum component flux =
AIPS 1: (source * beam)
AIPS 1: BPARM .3 10 (1) Inner region radius (deg)
AIPS 1: 0 1.2 (2) Field overlap (pixels)
AIPS 1: 5.000E-04 256 (3) Factor to scale NVSS
AIPS 1: *rest 0 fluxes, 0 -> 1
AIPS 1: (4) Radius NVSS search (deg)
AIPS 1: (5) Flux limit in NVSS (Jy)
AIPS 1: (6) IMSIZE for NVSS fields
AIPS 1: (7) IMSIZE for Sun fields
AIPS 1: (8) Write Clean boxes for
AIPS 1: NVSS fields
AIPS 1: (9) Maximum allowed phase
AIPS 1: error in imaging
AIPS 1: (10) Points per beaam
AIPS 1: PBPARM *all 0 Beam parameters:
AIPS 1: (1) Cutoff; (2) Use (3)-(7)
AIPS 1: (3)-(7) Beam shape parms
AIPS 1: NVSS input file name
AIPS 1: ' ' => AIPS provided.

This tiles 7 "big" (1024x1024) fields in the center of my image (within 0.3 degrees).
And then it puts smaller (256x256) fields on any NVSS sources with flux greater than 5e-4 Jy which are located between 0.3 and 1.2 degrees of the center of my field. In this case, that gave 31 external fields.

As Joe Lazio suggests, you might want to do a quick imaging run of all these external fields to see which ones actually contain sources, and delete from the BOX file any fields that don't really contain much of anything in your images. If you do delete some fields, you'll have to renumber the renaming ones or I think IMAGR will balk at you.

When you're ready to image with IMAGR, here are a few important things to keep in mind:
IMSIZE-- is the minimum field size! So in the above case, I'd set it to 256. IMAGR is smart enough to know to make the bigger internal fields 1024x1024 (because the BOX file tells it to).
NFIELD-- set this to the total number of fields in your BOX file.
BOXFILE-- set this to the file that came out of SETFC
OBOXFILE-- IMAGR will be smart enough to print out the field locations/sizes, AND any clean boxes to this file. Hurrah!

When you do start imaging, if you have DOTV = 1, IMAGR will show you one field at a time, and I think it tries to show you which ever field has the most flux in it (or the most flux remaining in it, after cleaning.)


emily said...
This comment has been removed by the author.
Anonymous said...

I'm new to AIPS, and noticed that using SETFC automatically sets clean boxes for every facet, the same size as every facet. When I've been wide field imaging, I've just been adding smaller clean boxes around sources on top of these. Is this ok, or is it better to delete these automatic boxes, and just manually add ones around obvious sources? Thanks.

Anonymous said...

can someone help me out by writing the detail process of 3d imaging/cleaning after split file is created? I am new to AIPS and any kind of help is welcomed. thanks in advance..