On 17 Jan 2012, at 16:36, Eric Greisen wrote:

> Andy Biggs wrote:
>> Hi. I'm using FLATN to combine 4 separate GMRT 610-MHz pointings together. It is generally acting sensibly, but there is a strange pattern around the main area of interest that I don't think should be there. Please see attached figure. Note that each pointing is very large and should be greatly truncated by FLATN i.e. the pixels with the holes in the centre probably correspond to real data.
>> I am using the recommended parameters for the GMRT beam and I thought at first that the problem might be due to the GMRT polynomial increasing back above my cutoff of 0.1 at large distances. However, I just tried it again with PBPARM(6)=0 and that shows the same problem.
>> Are my inputs okay?
> 
> I agree that this looks wierd - it is as if outside some ellipse it allows values from the outer facets even though one would hope that they would be blanked.
> 
> PBPARM(1)=0 and PBPARM(1) = 0.1 should do the same thing according to the help file.  Try PBPARM(1) = 0.02 and 0.2 instead and maybe I can get some idea what is going on - it would help to see an IMHEADER of an outer facet of say the top left pointing plus the IMHEAD of the center facet.
> 
> Curious - BTW a pixel is a single point like your bright "dots" in the good part.  Of course the worry is that the good part includes outer facet messed up from the more distant pointings...
> 
> I will start looking although I have been told to write a very long and tricky document so software is taking a back seat.
> 
> Eric Greisen