Difference between revisions of "Image"

From CASA Guides
Jump to: navigation, search
(Created page with '{{Simulations Intro}} Category: Simulations * <tt>simdata</tt> just calls a subset of<tt>clean</tt>, so you should refer to help clean for more information. * '''vis'''…')
 
Line 18: Line 18:
 
* <tt>simdata</tt>'s default clean parameters are not particularly aggressive and almost certainly not optimal for your application. I recommend that you run <tt>simdata</tt> and then clean the measurement set in various ways to determine the best results.  Once a deconvolved image exists, as long as you name it $project.image, you can run  
 
* <tt>simdata</tt>'s default clean parameters are not particularly aggressive and almost certainly not optimal for your application. I recommend that you run <tt>simdata</tt> and then clean the measurement set in various ways to determine the best results.  Once a deconvolved image exists, as long as you name it $project.image, you can run  
 
only the <tt>analyze</tt> subtask of <tt>simdata2</tt> to calculate difference and fidelity images.
 
only the <tt>analyze</tt> subtask of <tt>simdata2</tt> to calculate difference and fidelity images.
 +
 +
* <font color="green">Single-dish only NOTE</font>:  If you are only simulating a single dish observation, <tt>simdata</tt> does not perform deconvolution, so '''niter''', etc are not relevant.  The useful parameters are restricted to '''vis''', '''imsize''', '''cell''', and '''stokes'''.
 +
  
 
== Graphic Output ==  
 
== Graphic Output ==  

Revision as of 15:12, 9 June 2010

Simulating Observations in CASA

  • simdata just calls a subset ofclean, so you should refer to help clean for more information.
  • vis can be one Measurement Set, or a list e.g. "[project.ms,project.sd.ms]" if you want to deconvolve both Single Dish and interferometric data into images. simdata will first (or only, if you're doing total power only) grid the single dish data, then use the resulting image as modelimage input to the clean task. Various permutations of this are possible. If you run simdata once only in single dish mode, and create a gridded image, then you can return later and run simdata again predicting interferometric visibilities, and either specify the single dish measurement set in vis, or only specify the interferometric measurement set there, and specify the existing single dish image directly in modelimage.
  • imsize and cell specify the output image just as in clean
  • Cotton-Schwab clean is selected for single fields, and Mosaic gridding for mosaic simulations. For more detail and flexibility see the clean task.
  • weighting selects natural, briggs (robust=0.5), or uniform weighting.
  • outertaper allows application of an additional uv taper to the outer scales.
  • niter and threshold specify how deeply to clean (niter=0 is acceptable to create a dirty image)
  • simdata's default clean parameters are not particularly aggressive and almost certainly not optimal for your application. I recommend that you run simdata and then clean the measurement set in various ways to determine the best results. Once a deconvolved image exists, as long as you name it $project.image, you can run

only the analyze subtask of simdata2 to calculate difference and fidelity images.

  • Single-dish only NOTE: If you are only simulating a single dish observation, simdata does not perform deconvolution, so niter, etc are not relevant. The useful parameters are restricted to vis, imsize, cell, and stokes.


Graphic Output

If you have graphics turned on, you will see output such as below:

N891d.image.png