[daip] Data combination

[Eric Greisen]

Danielle Fenech writes:

 Please use daip at nrao.edu - I do go on vacation at times including
next week.

 > 
 > I'm trying to combine two sets of different frequency data within aips and 
 > require them to be combined but retain the original frequency information 
 > for each dataset.
 > 
 > Thus far, I have tried to combine the datasets, using dbcon, as different 
 > subarrays, which seems to work as required, shifting the u,v,w values 
 > for the second frequency data to the appropriate reference channel. 
 > However, if I then try to image/split and use the combined data, it is 
 > calculating the u,v,w values from the header reference frequency and 
 > ignoring the required frequency information (stored in the antenna table 
 > for the appropriate subarray). Is there any way I can convince imagr/split 
 > and the like to check for the reference frequency information in the 
 > antenna table first?
 > 
 > Otherwise, do you know of any other way to combine the datasets and retain 
 > the appropriate frequency information?
 > 
 > I don't know if anyone has tried a similar thing before but I hope that 
 > made some sense and any suggestions/ideas would be much appreciated.

It is done all the time.  But I do not understand what you are trying
to do - imagr and split are very different tasks so what you mean by
"imagr/split" is beyond me.  IMAGR does use the correct u,v,w as
recorded with the data to make the images and can handle different
frequency increments in different IFs.  It does assume that the
frequency increment is the same in separate subarrays and will not
scale the u,v,w of multiple channels entirely correctly for subarrays
> 1 when doing bandwidth synthesis (many channels into one image).

SPLIT is used to apply calibration which I would assume you have
already done before DBCON.  Even when doing self-cal, I would
recommend self-scaling each data set separately and then combining
them again, for the same scaling of multiple channels issues.

Eric Greisen