[daip] Frequency coordinates(31DEC05 vs 31DEC04)

[Eric Greisen]

Neeraj Gupta writes:
 > Dear Eric,
 > 
 >       Thanks a lot for the message.
 > 
 > > there will be many other differences.  Are you claiming that on the
 > > very same data set, IMAGR now does the "wrong" thing whereas it used
 > > to do the "correct" thing (i.e. ignore the FQ table)?
 > 
 >        Yes, that is exactly the point.  The illustration that I appended to 
 > the last mail used same dataset through 31DEC04 and 31DEC05.
 > 

I have attempted to test this - only briefly yesterday.  I found
puzzling results.  I forced a value of -10**6 for the increment in the
header and in the FQ table left the +10**3 in the FQ table but changed
the sideband code to -1.  Running in OLD (04), NEW (05) and TST (06)
IMAGR gave a cube with increment -10**6 on both OLD and NEW and -10**3
in TST.  This is very contrary to my belief that NEW and TST should
not differ and contrary to your finding that OLD and NEW do differ.

I will explore....

Are you running from a multisource data set or a single source data
set?  This might alter the behavior.

Eric Greisen