[daip] fshift

Eric Greisen egreisen at nrao.edu
Thu Apr 4 17:16:33 EDT 2013


Katherine Blundell wrote:
> Hi Eric,
> 
> Sorry not to have been in touch for ages - much busyness here in recent weeks.   I'm afraid that FSHIFT, that you kindly put into JMFIT, is still puzzling us.   
> 
>>From a 7x7 box centred on the brightest pixel, the following is pasted from JMFIT outputs:
>   X-position        =   1047.835 +/-   0.0144 pixels
>   Y-position        =    955.902 +/-   0.0155 pixels
>                       RA 19 11 49.5642089  +/-  0.000001442
>                       DEC 04 58 57.769581  +/-   0.00002320
>   RASHIFT=   0.000269 DECSHIFT=   0.000081 to center on pixel
> 
> But, I calculate that from the X-position and Y-position found above,  that rashift and decshift should be as follows (N.B. cellsize = 0.0015,  and the Dec of the target (SS433) is 4.9825 degrees), typing everything into python at the command line.
> 
> for RASHIFT:
>>>> (1048.0-1047.835)*0.0015/numpy.cos(3.141592*4.9825/180)
> 0.00024843878223426024
> 
> for DECSHIFT:
>>>> (956.0-955.902)*0.0015
> 0.0001469999999999345
> 
> So, RASHIFT seems to be slightly different and DECSHIFT is out by a factor of nearly 2.     Please accept my apologies if I am doing something silly, but I can't see where AIPS is getting the FSHIFT values from.   I'd be immensely grateful for your help in troubleshooting this.
> 
> Do hope that this finds you well?

Yes finally - I just had a trip to India (Assam and Aranuchal Pradesh) 
in which I had a terrible cold when I arrived in Delhi and the 2 days I 
had to sleep it off were not enough.  From then we walked and walked and 
walked and ...

But we saw cool things including a prolonged view of a very large female 
tiger as basically the last sighting of the trip!  Many not so good 
photos (it was getting dark).

I fear the aips implementation is not correct in that I multiplied by 
cos(dec) in the FSHIFT(1).  In fact it should just be delta pixel * 
pixel spacing for both FSHIFTs.  I will fix.  BUT, aips also adds any 
pre-existing shift and you did not tell me what IMHEADER would say about 
that...

Cheers

Eric





More information about the Daip mailing list