NRAO Home > CASA > CASA Task Reference Manual

0.1.145 widefield

Requires:

Synopsis
Wide-field imaging and deconvolution with selected algorithm

Description

This is the main wide-field imaging/deconvolution task. It uses the wprojection method for a large field of view, can make many facets, and can include outlier fields. Several deconvolution algorithms are supported. Interactive cleaning is also supported

Arguments





Inputs

vis

name of input visibility file

allowed:

stringArray

Default:

imagename

Pre-name of output images

allowed:

any

Default:

variant

outlierfile

Text file with image names, sizes, centers

allowed:

string

Default:

field

Field Name

allowed:

string

Default:

spw

Spectral windows:channels: ” is all

allowed:

any

Default:

variant

selectdata

Other data selection parameters

allowed:

bool

Default:

False

timerange

Range of time to select from data

allowed:

string

Default:

uvrange

Select data within uvrange

allowed:

string

Default:

antenna

Select data based on antenna/baseline

allowed:

string

Default:

scan

scan number range

allowed:

string

Default:

mode

Type of selection (mfs, channel, velocity, frequency)

allowed:

string

Default:

mfs

niter

Maximum number of iterations

allowed:

int

Default:

500

gain

Loop gain for cleaning

allowed:

double

Default:

0.1

threshold

Flux level to stop cleaning. Must include units

allowed:

any

Default:

variant 0.0Jy

psfmode

Algorithm to use (clark, hogbom)

allowed:

string

Default:

clark

ftmachine

Gridding method for the image (wproject, ft)

allowed:

string

Default:

facets

Number of facets along each axis in main image only

allowed:

int

Default:

3

wprojplanes

Number of planes to use in wprojection convolutiuon function

allowed:

int

Default:

64

multiscale

set deconvolution scales (pixels), default: multiscale=[]

allowed:

intArray

Default:

negcomponent

Stop cleaning if the largest scale finds this number of neg components

allowed:

int

Default:

0

interactive

use interactive clean (with GUI viewer)

allowed:

bool

Default:

False

mask

cleanbox(es), mask image(s), and/or region(s)

allowed:

any

Default:

variant

nchan

Number of channels (planes) in output image

allowed:

int

Default:

1

start

First channel in input to use

allowed:

any

Default:

variant 0

width

Number of input channels to average

allowed:

any

Default:

variant 1

imsize

Image size in pixels (nx,ny), single value okay

allowed:

intArray

Default:

256256

cell

arcsec

The image cell size in arcseconds [x,y], single value okay.

allowed:

doubleArrayarcsec

Default:

1.01.0

phasecenter

Field Identififier or direction of the image phase center

allowed:

any

Default:

variant

restfreq

rest frequency to assign to image (see help)

allowed:

string

Default:

stokes

Stokes params to image (I,IV,QU,IQUV,RR,LL,XX,YY,RRLL,XXYY)

allowed:

string

Default:

I

weighting

Weighting to apply to visibilities

allowed:

string

Default:

natural

robust

Briggs robustness parameter

allowed:

double

Default:

0.0

npixels

number of pixels to determine cell size for superuniform or briggs weighting

allowed:

int

Default:

0

noise

noise parameter for briggs abs mode weighting

allowed:

any

Default:

variant 1.0Jy

cyclefactor

Threshold for minor/major cycles (see pdoc)

allowed:

double

Default:

1.5

cyclespeedup

Cycle threshold doubles in this number of iterations

allowed:

int

Default:

-1

npercycle

Number of iterations before interactive masking prompt

allowed:

int

Default:

100

uvtaper

Apply additional uv tapering of visibilities.

allowed:

bool

Default:

False

outertaper

uv-taper on outer baselines in uv-plane

allowed:

stringArray

Default:

innertaper

uv-taper in center of uv-plane

allowed:

stringArray

Default:

1.0

restoringbeam

Output Gaussian restoring beam for CLEAN image

allowed:

stringArray

Default:

calready

Create scratch columns and store model visibilities so that selfcal can be run after clean

allowed:

bool

Default:

False

Returns
void

Example

 
 
        Wide-field imaging and deconvolution with selected algorithm:  
 
        This is the main wide-field imaging/deconvolution task.  It  
        uses the wprojection method for a large field of view, can  
        make many facets, and can include outlier fields.  Several  
        deconvolution algorithms are supported.  Interactive cleaning  
        is also supported.  
 
For making large images (>2000 on a size), see hints at the  
        end of the descriptions.  For making images larger than about  
        5000x5000, the available memory must be larger than 2 Gbytes. For such  
        images therefore  a computer with a 64-bit operating system may be  
        needed.  
 
 
        Keyword arguments:  
        vis -- Name of all input visibility files  
                default: none; example: vis=’ngc5921.ms’  
                example: vis=[’data01.ms’, ’data02.ms’]  
        imagename -- Pre-name of output images:  
                default: none; example: imagename=’n5921’  
                if outlier fields are included, then  
                   imagename=[’n5921’, ’outlier1’, outlier2’]  
                   and the first imagename is the wide-field image  
                output images names are: n5921.clean, n5921.residual,  
                n5921.model, n5921.interactive.mask  
        mode -- Type of selection  
                default: ’mfs’; example: mode=’channel’;  
                Options: ’mfs’, channel, velocity, frequency’  
        alg -- Algorithm to use  
                default: ’clark’;  
                Options: ’clark’, ’hogbom’,’multiscale’,’entropy’  
                    Strongly advise ’clark’.  multiscale and entropy  
                    well-tested.  
        imsize -- Image pixel size (x,y)  
                default = [256,256]; example: imsize=[500,500], or imsize=500  
                example for multiple fields: imsize=[(1000, 1000), (100, 100)]  
        cell -- Cell size (x,y)  
                default=[’1arcsec,’1arcsec’]  
                example: cell=[’0.5arcsec,’0.5arcsec’], or cell=’0.5arcsec’  
        phasecenter -- direction position or the field for the image center  
                A list of the above is needed for multiple-fields  
                default: ’’ -->field=’0’ as center; example: phasecenter=’6’  
                   phasecenter=’J2000 19h30m00 -40d00m00’  
                   phasecenter=[’J2000 19h30m00 -40d00m00’, ’J2000 19h57m00 40d00m00’]  
                      for wide-field, plus one outlier field.  
        stokes -- Stokes parameters to image  
                default=’I’; example: stokes=’IQUV’;  
                Options: ’I’,’IV’,’IQU’,’IQUV’  
        niter -- Number iterations, set to zero for no CLEANing  
                default: 500; example: niter=500  
        gain -- Loop gain for CLEANing  
                default: 0.1; example: gain=0.1  
        threshold -- Flux level at which to stop CLEANing (units=mJy)  
                default: 0.0; example: threshold=0.0  
        mask -- Name(s) of mask image(s) used for CLEANing  
                default: ’’  example: mask=’orion.mask’  
                Number of mask fields must equal number of imaged fields  
        cleanbox -- List of [blc-x,blc-y,trc-x,trc-y] values  
                default: []; example: cleanbox=[110,110,150,145]  
                Note: This can also be a filename with clean values:  
                fieldindex blc-x blc-y trc-x trc-y  
                cleanbox = ’interactive’ is very useful.  
        --- Data Selection  
        nchan -- Number of channels to select  
                default: 1; example: nchan=45  
        start -- Start channel, 0-relative  
                default=0; example: start=5  
                if mode=’frequency’ then a frequency value e.g start=’1.4GHz’  
        width -- Channel width (value > 1 indicates channel averaging)  
                default=1; example: width=5  
                if mode=’frequency’ then a frequency value e.g  width=’10kHz’  
        step -- Step in channel number  
                default=1; example: step=2  
        field -- Select field using field id(s) or field name(s).  
                  [run listobs to obtain the list id’s or names]  
               default: ’’=all fields  
               If field string is a non-negative integer, it is assumed a field index  
                 otherwise, it is assumed a field name  
               field=’0~2’; field ids 0,1,2  
               field=’0,4,5~7’; field ids 0,4,5,6,7  
               field=’3C286,3C295’; field named 3C286 adn 3C295  
               field = ’3,4C*’; field id 3, all names starting with 4C  
               example for multiple ms in vis parameter:  
               field=[’0~2’, ’1,2’]  
        spw -- Select spectral window/channels  
               default: ’’=all spectral windows and channels  
               spw=’0~2,4’; spectral windows 0,1,2,4 (all channels)  
               spw=’<2’;  spectral windows less than 2 (i.e. 0,1)  
               spw=’0:5~61’; spw 0, channels 5 to 61  
               spw=’0,10,3:3~45’; spw 0,10 all channels, spw 3, channels 3 to 45.  
               spw=’0~2:2~6’; spw 0,1,2 with channels 2 through 6 in each.  
               spw=’0:0~10;15~60’; spectral window 0 with channels 0-10,15-60  
               spw=’0:0~10,1:20~30,2:1;2;3’; spw 0, channels 0-10,  
                        spw 1, channels 20-30, and spw 2, channels, 1,2 and 3  
               For multiple ms in vis parameter:  
               spw=[’0,10,3:3~45’, ’<2’]  
        timerange -- Select time range subset of data (not implemented yet)  
            default=’’ meaning no time selection  
            example: timerange=’YYYY/MM/DD/HH:MM:SS.sss’  
            timerange=’< YYYY/MM/DD/HH:MM:SS.sss’  
            timerange=’> YYYY/MM/DD/HH:MM:SS.sss’  
            timerange=’ddd/HH:MM:SS.sss’  
            timerange=’< ddd/HH:MM:SS.sss’  
            timerange=’> ddd/HH:MM:SS.sss’  
        restfreq -- Specify rest frequency to use for image  
            default=’’ (i.e., try to use the restfreq specified in the visibility data)  
 
        --- Weighting  
        weighting -- Weighting to apply to visibilities  
                default=’natural’; example: weighting=’uniform’;  
                Options: ’natural’,’uniform’,’briggs’,’briggsabs’,’radial’, ’superuniform’  
        robust -- ’briggs’ and ’brigssabs’ robustness parameter  
                default=0.0; example: robust=0.5;  
                Options: -2.0 to 2.0; -2 (uniform)/+2 (natural)  
        npixels -- number of pixels to determine uv-cell size for weight calculation  
                 -- Used with superuniform or briggs weighting schemes  
                  example: npixels=3  
 
        --- widefield controls  
        ftmachine -- Gridding method for the image;  
                ft (standard interferometric gridding).  
                wproject (wprojection algorithm for gridding)  
                default: wproject  
        wprojplanes -- Number w-projection planes to use for gridding  
                default: 256  
                example: wprojplanes=64  
   Good value = BMAX(klambda) * Map width(arcmin)^2 / 600  
        facets   -- Number of facets along one axis on central image  
                image is divided in facets x facets rectangles.  
                default: 1  
                example: facets=3 makes 3x3 images to cover the field  
if ftmachine = ’ft’, only faceting is used  
                if ftmachine = ’wproject’, both wplanes and faceting  
                         can be used  (see below).  
 
        cyclefactor -- Change the threshold at which the deconvolution cycle will  
                stop and degrid and subtract from the visibilities. For bad PSFs,  
                reconcile often (cyclefactor=4 or 5); For good PSFs, use  
                cyclefactor 1.5 to 2.0.  
                default: 2.5; example: cyclefactor=4, but decreases speed considerably.  
                <cycle threshold = cyclefactor * max sidelobe * max residual>  
        cyclespeedup -- Cycle threshold doubles in this number of iterations  
                default: -1; example: cyclespeedup=500  
 
        --- MEM parameters (Experimental, not well-tested)  
        sigma -- Target image sigma  
                default: ’0.001Jy’; example: sigma=’0.1Jy’  
        targetflux -- Target flux for final image  
                default: ’1.0Jy’; example: targetflux=’200Jy’  
        constrainflux -- Constrain image to match target flux;  
                otherwise, targetflux is used to initialize model only.  
                default: False; example: constrainflux=True  
        prior -- Name of MEM prior images  
                default: [’’]; example: prior=’source_mem.image’  
 
        --- Multi-scale parameters (Experimental, not well-tested)  
        negcomponent -- Stop component search when the largest scale has found this  
                number of negative components; -1 means continue component search  
                even if the largest component is negative.  
                default: 2; example: negcomponent=-1  
        scales -- Used for alg=’multiscale’; set a number of scales or a vector  
                default: [0,3,10]; example: scales=[0.0,3.0,10.0, 30]  
        --  interactive masking  
        npercycle -- when cleanbox is set to ’interactive’,  
           this is the number of iterations between each clean to update mask  
           interactively. However, this number can be adjusted during execution.  
 
uvtaper -- Apply additional uv tapering of the visibilities.  
               default: uvtaper=False; example: uvtaper=True  
                  uvtaper=True expandable parameters  
                     outertaper -- uv-taper on outer baselines in uv-plane  
                           [bmaj, bmin, bpa] taper Gaussian scale in uv or  
                            angular units. NOTE: uv taper in (klambda) is  
                            roughly on-sky FWHM(arcsec/200)  
                         default: outertaper=[]; no outer taper applied  
                            example: outertaper=[’5klambda’]  circular taper  
                                FWHM=5 kilo-lambda  
                                outertaper=[’5klambda’,’3klambda’,’45.0deg’]  
                                outertaper=[’10arcsec’] on-sky FWHM 10"  
                                outertaper=[’300.0’] default units are meters  
                                   in aperture plane  
                     innertaper -- uv-taper in center of uv-plane  
                             NOT YET IMPLEMENTED  
 
        restoringbeam -- Output Gaussian restoring beam for CLEAN image  
                [bmaj, bmin, bpa] elliptical Gaussian restoring beam  
                default units are in arc-seconds for bmaj,bmin, degrees  
                for bpa default: restoringbeam=[]; Use PSF calculated  
                from dirty beam.  
                example: restoringbeam=[’10arcsec’] circular Gaussian  
                       FWHM 10" example:  
                       restoringbeam=[’10.0’,’5.0’,’45.0deg’] 10"x5"  
                       at 45 degrees  
 
calready -- if True will create scratch columns if they are  
                not there. And after clean completes the predicted model  
                visibility is from the clean components are  
                written to the ms.  
 
        async --  Run asynchronously  
                default = False; do not run asychronously  
 
 ======================================================================  
 
                      HINTS ON RUNNING WIDEFIELD  
 
      1.  Decide if the images will be specified directly in the  
          inputs or with an outlier file.  For more than a few fields,  
          an outlier file more convenient.  
 
         Direct Method:  
 
            cell = [’1.0arcsec’, ’1.0arcsec’]  
            imagename = [’M1_0’,’M1_1’,’M1_2]  
            imsize = [[1024,1024],[128,128],[128,128]]  
            phasecenter = [’J2000 13h27m20.98 43d26m28.0’,  
                      ’J2000 13h30m52.159 43d23m08.02’, ’J2000 13h24m08.16 43d09m48.0’]  
 
          Text file method  (in outlier.txt)  
 
            imagename = ’M1’  
            outlierfile = ’outlier.txt’  
               [phasecenter, imsize ignored]  
 
            Contents of outlier.txt  
            C   0   1024 1024   13 27 20.98     43 26 28.0  
            C   1    128  128   13 30 52.158    43 23 08.00  
            C   2    128  128   13 24 08.163    43 09 48.00  
 
         In both cases the following images will be made:  
             M1_0.image, M1_1.image, M1_2.image     cleaned images  
             M1.0.model, M1_1.model, M1_2.model     model images  
             M1.0.residual, M1_1.residual, M1_2.residual     residual images  
 
       2.  Wprojection:  It is fastest to use wprojection without faceting.  
             ftmachine = ’wproject’  
             wprojplane = NN  
 
           The value of NN should be chosen as small as possible to reduce  
           execution time.  The algorithm  
               NN = BMAX(klambda) * imagewidth (arcmin)^2 / 600, with a minimum  
                    of 16, should be adequate.  
 
       3.  Depending on the memory of the computer, a limit of about  
       5000x5000 may occur for example if a computer has 2Gbyte of  
       RAM. Also a 32-bit computer has a maximum limit of 2Gbyte  
       memory usable per process, irrespective of how much physical  
       RAM is present. Hence it is recommended to move to a 64-bit  
       computer with more than 2 GByte of RAM for >5000x5000 images  
 
 
       4. For data with extremely large ’w’ values, i.e low frequency,  
       long baseline and very widefield image, the wprojection  
       convolution can be very large and either not fit in memory or  
       slow for processing.  In these cases you should consider using  
       both ftmachine=’wproject’ and facets=xx where is 3.  
 

__________________________________________________________________

Permissions


More information about CASA may be found at the CASA web page

Copyright © 2016 Associated Universities Inc., Washington, D.C.

This code is available under the terms of the GNU General Public Lincense


Home | Contact Us | Directories | Site Map | Help | Privacy Policy | Search