First Steps to Visibility Forward Fitting

• Select a flare time interval, xyoffset, and energy band
• Compute visibilities using hsi_vis_usershell.pro;
  
  • We'll use the June 17, 2003 flare as an example:
    

         time_range = '2003-jun-17 '+['22:53:10','22:54:40']
         energy_band = [40.,60.]
         xyoffset = [-790.,-140.]
         vis=hsi_vis_usershell(time_range=time_range,xyoffset=xyoffset,energy_band=energy_band)
    

  • Edit visibilities (i.e. delete outliers)
    

         vise=hsi_vis_edit(vis)
    

  • Force visibilities to be conjugate (i.e. vis[-u,-v]=vis[u,v]* )
    

         visc=hsi_vis_combine(vise)
    

  • Optional: Display all the visibility amplitudes & phases in a multitude of windows:
    

         hsi_vis_display,visc
         wdel,/all              ; free up all the windows
    

  • Create a back projection map from the visibilities
    

         hsi_vis_bpmap,visc,time=time_range(0)
         hsi_vis_bpmap,visc,/over,/cont,lev=[10,20,30,40,50,60,70,80,90],/percent,c_color=1 ; overlay with contours
     

• Make a Clean, Mem_njit, or Pixon map to identify the starting parameters
• Choose a suitable model to match the maps (e.g. two circular Gaussians)
• Provide initial guesses for parameters (in this case srcfwhm, srcflux, srcx, srcx for two sources)
  

        srcstr={srctype:'circle',srcflux:0,srcx:0,srcy:0,srcfwhm:0,eccen:0.,srcpa:0.,loop_angle:0.,albedo_ratio:0.,srcheight:0.}
        srcstr=replicate(srcstr,2)  
        srcstr.srcflux=[48.,36.]      ; (These fluxes don't have to be accurate.)
        srcstr.srcx=[-798.,-782.]     ; (Approximate x positions.)
        srcstr.srcy=[-148.,-132.]     ; (And approximate y positions will do.)
        srcstr.srcfwhm=[3,3]          ; These FWHM values should be less than or roughly equal to the real widths
   

• Run hsi_vis_fwdfit to find the eight parameters:

        hsi_vis_fwdfit,visc,/multi,srcin=srcstr,srcout=srcfit,/noedit
        ; The /noedit option is required since we performed a vis_edit earlier.
  

• At this point the screen will display the eight parameters and their estimated sigmas.
• You should also check the error estimates of the fwdfit parameters by looking at the amplitudes
  • Overlay visibility amplitudes with amplitudes constructed from the forward fit
    

          window,1
          mapcenter = visc[0].xyoffset ; set the mapcenter to the vis phase center used earlier
          hsi_vis_fwdfit_plotfit,visc,srcfit,mapcenter,time=time_range(0)
      

  • Make a visibility bpmap

          window,2 
          hsi_vis_fwdfit_plotfit,visc,srcfit,mapcenter,/map_resid,time=time_range(0)
           
        

There are other tests, and various bells and whistles in the software, but soon these command-line utilities will be embedded in the objects and available from the gui, so this tutorial will become largely obsolete.