next up previous contents index
Next: IMAGE RECONSTRUCTION (DECONVOLUTION) Up: IMAGING Previous: Imaging multiple freqids and

Imaging spectral-line data

   

Compared with all this complicated continuum multi-frequency synthesis, spectral-line imaging is rather easy. You just need to make one image per channel, or perhaps one image per group of channels. UVMAP is not very useful for spectral line imaging, so I will just discuss HORUS\ and MX. The end result is a three-dimensional image, the third dimension being frequency. This is referred to as a cube. In spectral-line imaging, it is important that you do not make images at a frequency resolution that is much too great compared to the width of the features you are interested in. Otherwise you run the very real risk of not detecting your line. Make sure the frequency resolution is such that there are about 3-4 points across the line at most. This may require you to use the channel binning option of MX. If you are making a detection experiment, the best sensitivity to the putative line is obtained by matching the channel width to the line width.

A preceding chapter, §13 discusses other spectral-line issues, including continuum subtraction which you should do first.

  1. As usual, HORUS can be run on a multi-source file with the calibration applied directly (§ 12), or on a single-source file (any sort order) which has already had the calibration applied to it. The channel control is fairly rudimentary, and all you can do is give a channel start, end, and increment. As far as I am aware, the increment just skips channels, it doesn't average them together. However, you might like to apply some spectral smoothing with the smooth adverb. You can Hanning smooth by setting smooth=1,0 and then throw away every other channel with chinc=2.

    All other parameters should be as described in § 15.2.

    HORUS
    Apply calibration adverbs
    to multi-source files
    optype='line' Spectral line imaging of Stokes I
    smooth=1,0 Hanning smooth
    bchan=50 Image channels 50
    echan=450 to 450 at increments
    chinc=2 of 2 say

  2. For MX, you must first run SPLIT (§ 12) and possibly sort to XY order with UVSRT (§ 15.2) for very large images. MX provides more flexible channel selection than HORUS. I have described how the channel-selection adverbs work for MX in § 15.4.

    In the example below, pairs of channels are gridded together, starting at channel 50 and ending at channel 450. There is no smooth adverb in MX.

    MX
    in2name,in2class Name of visibility work file, leave blank
    in2seq,in2disk for first time, else fill in
    bif=2 Select desired IF
    eif=2
    bchan=50 Grid pairs of channels
    echan=450 from 50 to 450
    chinc=2 Skip every other channel in output
    npoints=2 Grid 2 channels together
    channel=0 No CLEAN restart
    nfield=1 Just one field for beginners
    fldsize=0 For CLEAN step only
    rashift=0 Don't shift image centre
    decshift=0 from phase centre
    niter=0 Don't CLEAN yet

Last update : 27/11/93



1


next up previous contents index
Next: IMAGE RECONSTRUCTION (DECONVOLUTION) Up: IMAGING Previous: Imaging multiple freqids and

nkilleen@atnf.csiro.au