Viewing Calibration/49473
Full headers

Private message: yes  no

Notes:

Notification:

Date: Mon, 23 Jan 2012 11:29:36 GMT
From: stefano.andreon@brera.inaf.it
To: xmmhelp@sciops.esa.int
Cc: stefano.andreon@brera.inaf.it
Subject: what PSF model in emldetect

Full_Name: Stefabi
Submission from: (NULL) (192.167.37.156)


Hi,

where may I found info about which PSF is used in emldetect to decide if a
source is extended (to compute the likelihood of extension)?  What was in SAS
6.5?
Of course, if a circular PSF is taken, the likelihood of extension is
overestimated
at large off axis angles.

Regards, Stefano Andreon

Reply 1

From: Matthias Ehle <xmmhelp@sciops.esa.int>
To: stefano.andreon@brera.inaf.it
Subject: Re: what PSF model in emldetect (PR#49473)
Date: Mon Jan 23 14:42:06 2012

Dear Stefano,

looking at the emldetect online documentation, e.g.
http://xmm.esac.esa.int/sas/current/doc/emldetect/node3.html

I read:

If parameter usecalpsf is set to true (this is the default) PSF fitting 
is performed using a tabulated energy and position dependent PSF as provided 
in the calibration database (accuracy level = medium). 
Alternatively, an internal, hardcoded PSF as specified in section ``Algorithm''

is used. The hardcoded PSF consists of a superposition of four Gaussian
profiles
with an off-axis dependent width. It is axially symmetric and currently has no 
energy dependence.

You can display the 'medium' accuracy PSFs with the SAS task 'calview'
(selecting
the Instrument, Accuracy Level, and commanding 'View -> PSF').
The PSF is part of the current calibration file (CCF, not coupled to a certain
SAS version). The most recent CCF is XRT#_XPSF_0013.CCF (# being 1,2,3 for the
three X-ray telescopes) released 2011-DEC-03. The update only affected the
ellbeta
accuracy level (currently not used by emldetect, as explained above), see
http://xmm2.esac.esa.int/external/xmm_sw_cal/calib/rel_notes/index.shtml

For your info, we are preparing the source detection with an improved 2-D PSF
model that is still under testing and expected to become functional with
the next SASv12 release.

I hope this answers your questions.

Regards,
   Matthias

--
   Dr Matthias Ehle 
   XMM-Newton User Support Group
   European Space Astronomy Centre

Followup 1

Date: Mon, 23 Jan 2012 16:00:21 +0100 (CET)
From: Stefano Andreon <stefano.andreon@brera.inaf.it>
To: Matthias Ehle <xmmhelp@sciops.esa.int>
Cc: stefano.andreon@brera.inaf.it
Subject: Re: what PSF model in emldetect (PR#49473)

Hi Matthias,
 	thank you for the quick reply. I miss a point, however:

On Mon, 23 Jan 2012, Matthias Ehle wrote:

> Dear Stefano,
>
> looking at the emldetect online documentation, e.g.
> http://xmm.esac.esa.int/sas/current/doc/emldetect/node3.html
>
> I read:
>
> If parameter usecalpsf is set to true (this is the default) PSF fitting
> is performed using a tabulated energy and position dependent PSF as
provided
> in the calibration database (accuracy level = medium).

Is such PSF azimuthal symmetric?

>
> You can display the 'medium' accuracy PSFs with the SAS task 'calview'
> (selecting
> the Instrument, Accuracy Level, and commanding 'View -> PSF').
> The PSF is part of the current calibration file (CCF, not coupled to a
certain
> SAS version). The most recent CCF is XRT#_XPSF_0013.CCF (# being 1,2,3 for
the
> three X-ray telescopes) released 2011-DEC-03. The update only affected the
> ellbeta
> accuracy level (currently not used by emldetect, as explained above), see
> http://xmm2.esac.esa.int/external/xmm_sw_cal/calib/rel_notes/index.shtml

I'm reading a paper that used SAS 6.5 and I would like to known, with 
that version, how much complex PSF the authors may have used to extimate 
the likelihood of extent: circular, elliptical, 2-component elliptical?

Regards, Stefano Andreon

Reply 2

From: Matthias Ehle <xmmhelp@sciops.esa.int>
To: stefano.andreon@brera.inaf.it
Subject: Re: what PSF model in emldetect (PR#49473)
Date: Mon Jan 23 15:51:37 2012

Hi Stefano,

you asked:
> Is such PSF azimuthal symmetric?
the "accuracy level=medium" PSF model is already not azimuthal symmetric.
It was based on a ray-tracing model and has angular dependencies included,
which you can see if you play around with the angles in calview
(the future 2D-PSF is based on actually observed point sources).

Cheers,
   Matthias