Theory of restoring

Is MIRIAD being a pain? Let us know your experience.

Moderator: Mark.Wieringa

Post Reply
Stephan
Posts: 7
Joined: Tue Mar 15, 2011 9:30 pm

Theory of restoring

Post by Stephan »

So, I'm now arriving at the imaging phase of my HI imaging project. I have a large set of different configurations. Some are 12 hours long, some only a couple. To add to the fun, there are multiple pointing offsets such that mosaicing is required. My main question is: how can I produced the 'best' restored image. I'm following the steps from the PhD thesis by Jess O'Brien. I'll use the reduction for ESO115-G021 as an example

First a general invert of the image
invert vis=../reductions/*/*.uvl map=first.map beam=first.beam options=mosaic,double robust=0.5 line=velocity,91,350,3.3,3.3 stokes=ii

Followed by a first mosmem run. I only set 5 iterations because I'm only interested in the beamsize from restor.
mosmem map=first.map beam=first.beam out=first.model region=quarter niters=5

Restor command. Not setting a fwhm or PA.
restor map=first.map beam=first.beam model=first.model out=first.restor

Using Gaussian beam fwhm of 18.975 by 10.043 arcsec.
Position angle: -55.5 degrees.


If I follow the way Jess did her reductions I would need to pick the major axis of the beam, so 18.975 arcseconds, and then use that as an input for the next phase of the invert. However, I learned that the restor beam is only based on the first file that goes into invert. In this case a rather short observation in the 1.5B configuration.

My question thus is: What is the recommended strategy in finding a restor beam size? Is there a way to find some optimal value between all these beams? Like an average or so? I'm aiming for the smallest beam possible. So do I have to place my longest 6* observation as the first input for invert?

Cheers,
Stephan
ste616
Site Admin
Posts: 220
Joined: Thu Feb 04, 2010 3:27 pm
Location: Paul Wild Observatory Narrabri NSW

Re: Theory of restoring

Post by ste616 »

Hi Stephan,

The invert task makes an image using all the available uv data, which in your case will contain many configurations. The dirty synthesised beam that invert makes will also be based on all the uv data.

The beam dataset from invert is the only thing restor knows about when you run it. It does a fit of this dirty beam to figure out how big to make the clean beam. So the beam size reported by restor will be the ideal beam size for the combination of configurations and weighting scheme that you have chosen. So you shouldn't have to figure the beam size out for yourself, as Miriad will do it properly, regardless of how many configurations you combine in invert, or which one you specify as the first file.

I hope this has helped, but please let us know if it's still not clear.
cheers
Jamie Stevens
ATCA Senior System Scientist
Post Reply