search

PaulHancock / MWA-fast-image-transients

Code and notes for the data reduction and analysis of fast image transients observed with the MWA
Other
5 stars 3 forks source link

Command line options for pixel and image size in im28s and im05s #27

Open gemmaanderson opened 5 years ago

gemmaanderson commented 5 years ago

Please make pixel and image size command line options for obs_im28s.sh and obs_im05s.sh as has been done for the obs_image.sh script.

Aesthetic option: change the name of the im28s scripts to im30s as the new rapid-response observations are now 120s (rather than 112s) in duration.

PaulHancock commented 5 years ago

The first part of this has been completed (the obs_*.sh scripts).

I haven't done the second part with the renaming.

PaulHancock commented 5 years ago

Are the new observations actually 120s of usable data, or are they still 120s observations with 112s of usable data?

gemmaanderson commented 5 years ago

That is a good question. I can't remember how to check actually.

gemmaanderson commented 5 years ago

For im28: I don't think that the image and pixel scales are being carried through properly. A test script has:

imsize=IMSIZE
pixscale=SCALE

There are no cleaning options (-c). Currently, niter is set to 0 Could you please include a similar cleaning setting as for normal imaging:

#if clean is set then change it to the default clean options.
if [[ ! -z ${clean} ]]
then 
    clean="-join-polarizations -niter 100000 -auto-mask 3 -auto-threshold 1 -circular-beam"
fi
PaulHancock commented 5 years ago

We previously decided that cleaning the 28s and 0.5s images wasn't going to be useful. It wasn't in the previous versions of the scripts, and so I didn't include it here. My reasoning was that these images are going to have a much lower SNR and either cleaning will not do anything useful (waste time) or it will pick non-source peaks and add false sources to the image (false transients).

We can revisit this if you feel that CLEAN is required.

gemmaanderson commented 5 years ago

While it makes sense to not clean the 0.5s as we are doing difference imaging and cleaning would result in more artefacts following a subtraction, it does make sense to clean at 28s, which we are not subtracting (currently).

I have just checked the header on a 28s image that I made of GRB 170827B, and they were definitely cleaned.

It may be best to include cleaning as an option in all of the imaging scripts so that people have the option if they choose.