Inserting a fake extended radio source

Is MIRIAD being a pain? Let us know your experience.

Moderator: Mark.Wieringa

saracmbr
Posts: 38
Joined: Fri Jan 20, 2012 3:40 pm

Inserting a fake extended radio source

Post by saracmbr »

Hello everyone,

I need to insert a fake extended radio source to one of my uv data which I already have reduced and calibrated. I don't know which task in Miriad I should use??
Do I need to use uvfit or imfit?
I need this analysis for my PhD thesis.
Can anyone help me get started?

Any help is appreciated.

Cheers,
Sara
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Inserting a fake extended radio source

Post by Mark.Wieringa »

Hi Sara,

Uvfit or uvsfit if you need spectral variation will do this.
Specify the source parameters with spar and object and make sure you specify fix for ALL the parameters. Specify the source flux as negative. Set the options to get out the residual data (and specify an output file).
When you run it uvfit complains there are no free parameters but subtracts off your source, giving a positive fake source in the data.

Cheers,

Mark
saracmbr
Posts: 38
Joined: Fri Jan 20, 2012 3:40 pm

Re: Inserting a fake extended radio source

Post by saracmbr »

Thanks Mark,

Do I need to use uvmodel afterwards to add the output file made by uvsfit to my actual uvdata or I just need to image the output of the uvsfit?

Cheers,
Sara
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Inserting a fake extended radio source

Post by Mark.Wieringa »

Hi Sara,

there is no need for uvmodel. You just give your calibrated data as the input for uvfit and the 'residual' will contain your original data plus the fake source. Then you just image that.
Uvmodel can be used if you have an image model of the source to simulate, but for simple gaussians uvfit can do the job.

Your simulated source will not have any calibration errors, no primary beam attenuation applied, etc., so it may not be directly comparable to real sources, but it can give a good feel for how your imaging and deconvolution process is recovering extended sources.

Cheers,

Mark
saracmbr
Posts: 38
Joined: Fri Jan 20, 2012 3:40 pm

Re: Inserting a fake extended radio source

Post by saracmbr »

Hi Mark,

I tried to use uvfit and uvsfit to model a point source first. I used exactly the same input arguments for both tasks. The image made after using uvfit has the same total flux I gave as input but the image made after using uvsfit has a random and different total flux from what I used as input. Are you sure that "uvsfit" is working correctly and it's not broken?

Thanks,
Sara
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Inserting a fake extended radio source

Post by Mark.Wieringa »

Hi Sara,

because uvsfit deals with spectral variation, you'll need to specify the spectral parameters sspar (alpha0,alpha1,alpha2) and also specify the sfix parameter as '012' otherwise it will try to solve for the spectral variation. Uvsfit should give the message that there are no free parameters if you are using it to simulate sources.

I have used uvsfit to put an extended source with spectral index in my data, and it comes out correctly.

Good luck,

Mark
saracmbr
Posts: 38
Joined: Fri Jan 20, 2012 3:40 pm

Re: Inserting a fake extended radio source

Post by saracmbr »

Hi Mark,

Do you know how I can eject a fake diffuse source with an arbitrary brightness profile (not a gaussian). I know the brightness profile equation.

Cheers,
Sara
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Inserting a fake extended radio source

Post by Mark.Wieringa »

Hi Sara,

the uvgen or uvsfit approach only works with the predefined source types (point, gaussian, disk, ring, shell). You can add several of these on top of each other to compose your source.
If you want something more general you'll need to create an image of the source and use uvmodel to calculate the visibilities. I think you should even be able to add a spectral index plane and it will be taken into account (you'd be simulating the output of mfclean).

As for creating the image, you could use imgen if it has the type of source you need, or produce it in some other package and import from fits.

Cheers,

Mark
saracmbr
Posts: 38
Joined: Fri Jan 20, 2012 3:40 pm

Re: Inserting a fake extended radio source

Post by saracmbr »

Hi Mark,

Do you think it is possible to change the "uvsfit" task original code to be able to generate visibilities with an arbitrary brightness profile? Is there ant way I could ask Miriad team to provide such modified task?

Cheers,
Sara
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Inserting a fake extended radio source

Post by Mark.Wieringa »

Hi Sara,

the way uvsfit and uvgen work is to specify the brightness profile analytically and calculate the visibility response from the (analytical) fourier transform of the brightness profile function. To use a arbitrary profile you need to use an image of the source and an FFT to calculate the visibility response. So, I think the answer to your question is 'no', I can't provide such a task for the generic case. If you have a specific request for a new functional form which has an analytical fourier transform I could try to add that.

The quickest route would be to generate an image of the source you want to simulate and then use uvmodel with options='add' to add it to an existing uv file.

Cheers,

Mark
Post Reply