[daip] [#4188]: blanking in GBT maps

[Eric Greisen]

Trisha wrote:
> Hi Eric,
> Thank you for your quick response.  I addressed some of your points below:
> 
> 
> On Mon, Dec 23, 2013 at 1:27 PM, Eric Greisen <do-not-reply at nrao.edu 
> <mailto:do-not-reply at nrao.edu>> wrote:
> 
>     There may be several issues here.  The blanked pixels are
>     created by SDGRD since you have not observed at every
>     position (usually) in the output image - otherwise you would be
>     dropping some data off the edges most likely.
> 
> My map is a raster scan so it should have data at every location that is 
> blanked (these blanked locations are occurring throughout the image).  
> Also, the blanked regions increase in number depending on the FLUX that 
> I set.  At a FLUX of -0.01 the map has no blanked regions.  So I do not 
> believe this is causing the problem.  

         FLUX has nothing to do with SDGRD.  I am trying to make sure 
you have a proper image cube first.  XMOM will work when that happens 
but not before.  You may have a regularly sampled region on the sky BUT 
the issue below means that the region you made an image for is almost 
certainly not neatly arranged on the sky.

> 
>     But I see another issue - the projection -GLS is not simple.
>     Reference declinations that are not zero cause the whole thing
>     to be gridded on an oblique projection that may be nothing like
>     you expected or wanted.   Frankly, I do not know what SDGRD
>     will now do with these projections and it would be good for me
>     to get my hands on a sample data set to try things out.  Perhaps
>     you could put a FITS file of the UV data somewhere I could find it.
> 
> The projection -GLS is something that I was simply told to use by the 
> GBT staff, therefore I have no feel for this causing a problem or not.  
> What I do know is that the data were taken in GLAT and GLONG and 
> apparently the output from GBTIDL is not compatible with AIPS (for the 
> coordinate system).  So in SDGRD I provide a fake center for my map as 
> if it was in RA and Dec, then, after SDGRD, I have to change the header 
> by force from RA and Dec to "GLAT-GLS" and "GLONG-GLS".  At least that 
> is my understanding from what I was told.  I can try to see if I can get 
> you a FITS file, however, would this cause a flux dependent blanking?

The advice you were given is obsolete.  GBT folks seem to insist on 
"rolling their own" and do not provide me any information whatsoever.
The aips task 'fittp' will write your file out to a disk file specified 
in dataout.  What I really want is the "UV" data file of single dish 
data not the image produced by SDGRD.

>  
> 
>     I your case, however, look at some of your channel images with
>     CBTR on the TV with DOCIRCLE true to see if the coordinates
>     are anything like what you expect.  Perhaps try SDGRD with
>     the -SIN projection which does not go into this oblique business.
> 
> I cannot find anything on CBTR, did you perhaps mean something else? 

I can't type well - try TVMOVIE and the task I suggested spelled 
correctly as CNTR.  I suspect that your SDGRD is not as you would have it.


>  
> 
>     In XMOM you want ICUT=0, FLUX=-1000000.
> 
> I have just tried this and the map remains blanked. 

XMOM will produce bad results until we get its input straight.

Eric Greisen