[daip] IMAGR question

[Eric Greisen]

On 09/21/2016 11:03 AM, Lisa Young wrote:
> Hi Eric.  I am hoping you can help me with a question on the way IMAGR
> has implemented the residual-rescaling method of Jorsater & van Moorsel
> 95.  I'm sure you remember that the issue is about the difference
> between the dirty beam area and the clean beam area, and that can be
> important when the clean components are added back to the residual image.
>

The code in AIPS is simple - it sums the beam over the specified 
rectangular area and the restoring beam over its Gaussian size.

I do NOT recommend this option.  If you image allowing IMAGR to fit the 
Clean beam to the dirty beam then you really do have the residual and 
components on a relatively close scale.  The actual scale is adjusted 
channel by channel so that all planes are on the same Jy per HEADER beam 
while the actual Clean beam parameters are stored in an attached (CG) 
table.  To put all planes of a cube on the same resolution, one then 
uses CONVL which will convolve each plane appropriately to get the 
desired output resolution.  This is then carefully consistent in the 
handling of units - the old approach is not.

Eric Greisen