Introduction to Miriad for CARMA users
A new miriad data-reduction carma-specific cookbook is now available: (as draft)
Usage Setup
From the Unix point of view MIRIAD is a set of programs that
you can execute from the command
line. So your Unix search path needs to be modified for this.
For most situations something like
source /usr/local/miriad/miriad_start.csh
will do, but check your local variations on the path name. For example, Maryland uses a
method referred to as astroload. At CARMA, depending if you are in the Owens Valley
or at Cedar Flat, you would use /sw/miriad/cvs/ or
/array/miriad/cvs/ resp.
Keywords
All MIRIAD programs need a supply of "keyword=value" pairs on the command line.
Often many take up a reasonable
default, but keywords such as the input file and the plotting device often do not have a
default.
- vis=: The name of the visibility dataset. MIRIAD datasets are always directories,
in which various files live (in MIRIAD often referred to as items)
- in=: A more generic name for an input dataset. Most image analysis programs use this.
- select=: Select certain visibilities for processing. It is a list of comma separated
actions, each of which is logically AND-ed. Example:
select=vis(20,30) would select visibilities 20 through 30.
select=ant(1)(3) would select baselines with only baseline 1-3.
select=ant(1,3) would select baselines with antennea 1 or 3.
- line=: Select how to re-sample the line channel data (this includes wideband). Examples:
line=channel,10,5,3,2 would select the 10 channels, starting at channel 5, in steps of 3, with width 2.
line=velocity,20,-100,10,20 would select 20 channels, starting at -100 km/s, in steps of 10, width 20 (km/s).
- region=: Selects a certain region (in ra-dec-vel) space to be used for image analyis.
Example:
region=box(20,20,30,30)(4,10) would select a region from pixel (20,20) to (30,30) and planes
4 through 10.
- device=: the pgplot plotting device. Commonly used are device=filename/ps for postscript,
and device=1/xs, device=2/xs for examples of persistent X-windows
The "miriad" shell
Some users, particularly if unfamiliar with a
program's inputs, prefer using MIRIAD via an AIPS-like shell called "miriad".
This is where you can select a task to run, inspect and set its parameters, and run the task.
Start the miriad shell by typing "miriad". Then you can
list a program's keywords and any values that are set by typing e.g. "inp uvplt".
You can assign values to the keywords with "keyword=value" (e.g.
vis=myfile.vis), check its inputs again by just typing "inp", see
explanations of the program and its keywords by typeing "help", and then
running it with "go". Keyword values are saved until overwritten or
explicitly unset. The command "mirhelp miriad" gives some more details.
Help!
The commands "doc" and "mirhelp" .... they explain all the keywords, their meaning and defaults.
Visibility Data Inspection
The data at CARMA are written in MIRIAD visibility format. Here are some frequently used
programs to inspect, visualize and calibrate these data.
- listobs: Makes a summary of a set of observations.
- varlist: List all uv variables in dataset
- uvlist: Print data, variables and statistics from uv dataset
- uvindex: Scan a uvdata file, and note when uvvariables change.
- uvplt: Make plots from a UV-data base on a PGPLOT device.
- varplt: Plot and list variables.
- uvspec: Plot averaged spectra of a visibility dataset.
Data Calibration and Manipulation
- gpplt: Plot and list gain and polarization correction terms.
- selfcal: Determine self-calibration of calibration gains.
- mfcal: Multifrequency antenna and passband calibration.
- uvcat: Catenate and copy uv datasets; Apply gains file, Select windows.
- uvcal: Apply special processing options to uv-data.
- uvaver: Copy and average a uv dataset.
Mapping
Once calibrated, conversion to an image (the FFT stage) and deconvolution can be done with the
following programs:
- invert: Transform multi-pointing visibility data into a map
- clean: Apply Hogbom, Clark or Steer CLEAN algorithm to a map
- mossdi: Apply Hogbom, Clark or Steer CLEAN algorithm to a map (mosaic'd version)
- restor: Restore clean components to make the CLEAN map
- maxen: Maximum Entropy deconvolution
- mosmem: Maximum Entropy deconvolution (mosaic'd version)
Image Data
Images (and cubes) can be inspected and visualized with the following tools:
- imhead: List items and pixel values from an image.
- histo: Find statistics of image and plot simple histogram
- cgdisp: displays and overlays images on a PGPLOT device
- imlist: List items and pixel values from an image
- imspec: calculate and plot map statistics
Starting with no data
If you have no data yet, and would still like to get started, two programs, depending on what flavor
of data you want to use, are useful:
uvgen: Visibility Data
This program needs a few small text files for the description of the source and the antenna configuration.
There are examples and templates in the $MIRCAT directory. Here is an example for continuum data:
% uvgen out=vis1 source=$MIRCAT/point.source ant=$MIRCAT/carma_de.04oct.ant telescop=carma
imgen: Image Data
This program creates maps or cubes, and is very easy to use. Example:
% imgen out=map1 object=gaussian spar=10,0,0,20,10,30
Installing
Some readers really want to know this ..... but installing MIRIAD
on your own machine is actually easy, assuming you have all the
right tools, compilers etc. installed. It is known to work out of the box on
most Linux and MacOSX machines.. On the simplest level it can be
done with the following two commands:
curl ftp://ftp.astro.umd.edu/progs/bima/miriad.tar.gz | tar zxf -
miriad/install/install.miriad
after which you can do
source miriad/miriad_start.csh
If that doesn't work, please refer to the
MIRIAD installation guidelines.
Page last modified:
Wednesday, 11-Jun-2008 12:50:34 EDT
by PJT.