[daip] Re: forwarded message from Eric Greisen

[Eric Greisen]

Bill Cotton writes:
 > Eric Greisen writes:
 >  > Rick gave me the paper with the L band pattern in it - I have now put
 >  > that in PBCOR et al for L and 2 of Rick's measured beams for P and 4.
 >  > That leaves the old ones for lambda < 15cm.
 >  > 
 >  > You did not answer the integerizing question - I know that OHGEO also
 >  > does the "skew" corrections for snap shots.
 >  > 
 > 
 >    Integerizing the reference pixel is (I think) generally (always?)
 > the wrong thing to do.  I believe that in our case the reference pixel
 > will be an integer in any case, so in a sense, it's not a problem.
 >    Another reason for not using HGEOM is that its interpolation
 > scheme isn't as good as the one in OHGEO when dealing with blanked
 > pixels.  HGEOM blanks anything even close to a blanked pixel.
 > 
 > -Bill

I have placed back new versions of HGEOM and OHGEO that default to the
center -> center when BOTH ref pixel adverbs are zero.  Errors with
BLC in OHGEO and with the definition of "center" in HGEOM and
integerizing in both tasks were corrected.

Sorry for the confusion.  I guess that this is as good a default as
any and if the user wants the ref pixel of image 2, the user can
specify it or use GETHEAD to read it to the adverbs.

Eric