[daip] Re: using DOALIGN in MCUBE

[Casey Law]

Hello Eric,
	Ok, I suspected this was the case.  Perhaps the help file could be
improved?  If you add the paragraph you just emailed me to the help file,
it would clear things up quite a bit.  For example, nowhere in the help
file does it say which coordinates it aligns; it simply refers to "the
coordinates".  Clarifying things like this would have saved me a week of
fiddling around and testing what this task was doing!
	I appreciate you taking the time to respond.  I just wanted to 
email my problem with the hopes that people don't run the risk of making 
the same mistakes in the future.

		casey

On Sun, 23 Nov 2003, Eric Greisen wrote:

> Casey Law writes:
>  > Hello,
>  > 	I am emailing you, since I read that you are the "Designated AIP".  
>  > I have a comment on the DOALIGN adverb and how it is used in MCUBE.  
>  > Either there is a bug here, or the wording of the help file could be
>  > improved (or I need to reread it...).
>  > 	The DOALIGN help file says that it: 
>  >  "Controls how two or more maps are aligned by various tasks." 
>  > It then describes how different values determine the various ways that it
>  > can align images.  However, when I feed MCUBE images which cover similar
>  > values of RA,Dec, but have different reference pixels and and RA,Dec
>  > values for their reference pixels, the alignment fails.  The result is a
>  > cube which has some of the images with their original coordinates, but
>  > some images have had a subim taken, and the coordinates are wrong.
>  > 	This problem occurs for DOALIGN=0.  My impression from the help
>  > file is that for this value of DOALIGN, it should recognize that the
>  > images do not have the same ref pixels and shift the input image
>  > accordingly.
>  > 	While I may have made a mistake in the above work, I have also
>  > heard from others that I shouldn't even try to use the DOALIGN adverb.  
>  > These people say that the *only* way to make a cube of properly aligned
>  > images is to use OHGEO prior to MCUBE.  But this impression from AIPS 
>  > users is contrary to what I read in the DOALIGN help file.  So I have two 
>  > questions:
>  >  1) Can DOALIGN meant to align images which share at least some RA,Dec
>  > values across the imaged region?  and
>  >  2) If yes, have there been any recent indications that it is working in 
>  > this way?
>  > 
> 
> The DOALIGN adverb only controls how images are lined up using their
> current pixel values.  They are not regridded.  OHGEO or HGEOM must be
> used if there are different reference value since the geometry is
> fundamentally different.
> 
> ERic Greisen
>