[daip] AIPS task UVMOD

[Eric Greisen]

On 04/26/2016 02:54 AM, lisakov wrote:
> Dear Eric,
>
> 1) Thank you for reply. Am I right in a guess that random numbers from
> -1 to 1 are distributed as a Gaussian?
>
> 2) I've checked RMS of original data (as I described before) and one of
> simulated data.  Then I've calculated the ratio of these RMSs and run
> UVMOD again with newFLUX = oldFLUX * ratio_of_RMS. And this gives
> intended result.
> So yet another question. Is there a way to deduce the input number of
> FLUX to get the same noise as of input data (without such a complex
> iterative way)?

I have changed the usage of FLUX when FACTOR=0 (causes input data to be 
ignored).  In that case the weights are meaningless, so FLUX is simply 
the noise in Jy from now on.  It retains the old meaning (better 
explained in the help file) when the input data are used (FACTOR not 0).

The random number generator produces an evenly distributed value from 0 
to 1.  But the routine that calls it does produce something like a 
Gaussian around 0 by summing the result 12 times starting at -6.

I hope