Offset Problem

[bnorfolk]

Hi,

I'm using the tasks invert, clean, restor, and cgdisp to image my target.

In doing this I've discovered my target is offset from the centre.

From reading a previous post I noticed the impos task was suggested but I was unsure on some points.

My general idea is that I want to use impos to find the offset of my source, input these values into invert, and then following clean, restor, and cgdisp to get an image.

But what do I input for the coord in the impos task, currently I have: impos in=myvis type="arcsec"

Thanks,
Brodie

[Mark.Wieringa]

Hi Brodie,

Firstly, it is not really a problem if your source is not in the centre, as long as your image is large enough to let you clean it properly.

If you do want to shift the field, the invert offset parameter can be use to specify the offsets in arcsec. As you suggest you can get these by reading off the pixel position of your source and translating it into arcsec offsets using impos (or by multiplying by the coordinate increments/cell size yourself).
If you have an absolute pixel position (e.g., 123,456) you would use impos in=myimage type=abspix coord=123,456 and then enter the offsets in arcsec into invert.
Take care with sign of the RA offset.

Cheers,

Mark

[ste616]

Hi Brodie,

I'm unsure on what you're after here. Are you trying to measure the flux density of your source, which is offset, or are you trying to make an image with the source at the centre?

If the latter, I wouldn't bother; there really should be no need to do that. But if you really do want to do that, be aware that by moving the source to the centre, you will likely run into trouble while correcting for the primary beam. You can shift the uv data to a different phase centre (which is what invert uses while imaging) using uvedit.

But if you only want to make a measurement of an off-centre source, image as normal, run the image through linmos to correct for the primary beam attenuation, measure your source offset with impos, then use those offsets as the initial guess for the x and y offsets in imfit. This is the method I would encourage.

[bnorfolk]

Thanks for the suggestions.