|
Introduction
This thread illustrates a few basic functionalities
of the LHEASOFT fitting
package XSPEC.
This thread is not intended to replace the full
XSPEC. Users
are encouraged to refer to the
XSPEC documentation
for any further information.
Expected Outcome
At the and of this thread the user will be able to:
- load all the files required by
XSPEC
- define and "fit" a model. We intend hereby by "spectral fitting" the
algorithm minimizing an appropriate goodness-of-fitness metrics by
varying the parameters of the model
- plot the data and the residuals against the best-fit model
SAS Tasks to be Used
None
Prerequisites
Either one of the following threads:
Useful Links
Caveats
|
Procedure
This thread contains a (very!) simple standard
XSPEC
session for fitting data from the XMM-Newton X-ray cameras. The EPIC case,
will be illustrated with an EPIC-pn spectrum (a fully analogous procedure
applies to EPIC-MOS spectra as well).
EPIC
We assume in the following that the following products have been created
and are available in the working directory: a source+background spectrum
(
PNspectrum.pi), a background only spectrum
(
PNspectrum_background.pi), a redistribution matrix (
PN.rmf),
and an ancillary file (
PN.arf), both of the later appropriate for
the source+background spectrum in use (see
Prerequisites for this
thread at the top of the page on how to create the spectra and associated
files for point-like sources with EPIC-pn or EPIC-MOS).
- Start XSPEC
xspec
- Load spectrum data:
data 1 PNspectrum.pi
- Load the background spectrum
backgrnd 1 PNspectrum_background.pi
- Load the redistribution matrix
response 1 PN.rmf
- Load the ancillary file
arf 1 PN.arf
- Create the output graphic plotting window
cpd /xs
- Define how data has to be plotted, e.g. in energy bins (wavelength would
be a valid alternative)
setplot energy
- Ignore bad channels in data
ignore bad
- Ignore energy intervals where the pn matrices are currently not calibrated
(e.g.: beyond the energy band 0.2-15.0 keV)
ignore **-0.2 15.-**
- Relax getting a look into your data
plot data
Fig.1: XSPEC data plot
- Load a model to be fitted (in this case a simple power law model with
photoelectric absorption)
model wabs*powerlaw
(you will have to hit return several times for every parameter of the model.
The following screen will appear:
mo = wa((po)).
Model: wabs[1]( powerlaw[2] )
Input parameter value, delta, min, bot, top, and max values for ...
Current: 1 0.001 0 0 1E+05 1E+06
wabs:nH>
Current: 1 0.01 -3 -2 9 10
powerlaw:PhoIndex>
Current: 1 0.01 0 0 1E+24 1E+24
powerlaw:norm>
---------------------------------------------------------------------------
---------------------------------------------------------------------------
Model: wabs[1]( powerlaw[2] )
Model Fit Model Component Parameter Unit Value
par par comp
1 1 1 wabs nH 10^22 1.000 +/- 0.
2 2 2 powerlaw PhoIndex 1.000 +/- 0.
3 3 2 powerlaw norm 1.000 +/- 0.
---------------------------------------------------------------------------
---------------------------------------------------------------------------
Chi-Squared = 2.7158111E+09 using 1675 PHA bins.
Reduced chi-squared = 1624289. for 1672 degrees of freedom
Null hypothesis probability = 0.
- Fit the parameters
fit 100 1e-1
The former number indicates the maximum number of iterations before the
minimization routine stops. The latter number indicate the minimum
difference in chi-squared between two minimization runs, before the
routine stops.
- Set the rebinning of your choice (for plotting purposes only!)
setplot rebin 3 4096
The former number indicates the maximum number of sigma to be accumulated
in a rebinned channel. The latter, the maximum number of channels to
be summed.
- Plot the data with the model adjustment and e.g. the residuals (alternatives,
ratio data/model, absolute chi-squared, etc)
plot data residuals
Fig.2: XSPEC data, model and residual plot
- Determine one-side errors on the best-fit parameters 1, 2 and 3
error 2.706 1 2 3
The first number indicates the delta chi-squared value.
RGS
We will assume in the following that products are available in the
work directory as
produced by SAS
rgsproc, the RGS SAS metatask for
whole data reduction (see
Prerequisites for this thread at the top of the
page on how to extract RGS spectra of a point-like source and associated
matrices):
- Start XSPEC
xspec
- Load the spectrum data (in this case the RGS1 background subtracted spectrum, first
order) into the first internal dataset
data 1:1 /xvsas05/sasval/data/procred/Mkn421/P0099280201R1S001SRSPEC1001.FIT
- Load the corresponding background file:
backgrnd 1 /xvsas05/sasval/data/procred/Mkn421/P0099280201R1S001BGSPEC1001.FIT
- Load the response matrix into the first internal response set. The SAS task
rgsrmfgen
puts both response and ancillary data into the response matrix
file
resp 1 /xvsas05/sasval/data/procred/Mkn421/P0099280201R1S001RSPMAT1001.FIT
- Loading a second dataset into second internal data position (in this case
data corresponding to RGS2)
data 2:2 /xvsas05/sasval/data/procred/Mkn421/P0099280201R2S002SRSPEC1001.FIT
- Load the corresponding background file:
backgrnd 2 /xvsas05/sasval/data/procred/Mkn421/P0099280201R2S002BGSPEC1001.FIT
- Load the corresponding response matrix
resp 2 /xvsas05/sasval/data/procred/Mkn421/P0099280201R2S002RSPMAT1001.FIT
A few differences need to be noticed at the data loading level
- rgsrmfgen (the SAS tasks to generate RGS response matrices,
which is by default automatically launched as the last stage of
rgsproc) creates a response matrix which includes both the
redistribution matrix and the effective area vector (ancillary file).
There is therefore no need for loading any *.arf files in
XSPEC when analyzing RGS spectra
- The syntax data n:n allows to keep the fit parameter
(fully or partly) independent between the two datasets during the
fitting procedure. The alternative syntax data n forces
the same model parameters to be fitted to all datasets
- Create the output graphic plotting window
cpd /xs
- Define how the data has to be plotted, e.g. in energy bins (wavelength would
be a valid alternative)
setplot energy
- Relax getting a look into your data
plot data
- Ignore bad channels in data
ignore bad
- Ignore energy intervals where the RGS matrices are currently not calibrated
(e.g.: beyond 1.9 keV)
ignore 1.9-**
- Load a model to be fitted (in this case a simple power law model)
model powerlaw
(you will have to hit return several times for every parameter of the model)
- Fit the parameters
fit 100 1e-1
The former number indicates the maximum number of iterations before the
minimization routine stops. The latter number indicates the minimum
difference in chi-squared between two minimization runs, before the
routine stops
- Set the rebinning of your choice (for plotting purposes only!)
setplot rebin 3 4096
The former number indicates the maximum number of sigma to be accumulated
in a rebinned channel. The latter, the maximum number of channels to
be summed.
- Plot the data with the model adjustment and e.g the residuals (alternatives,
ratio data/model, absolute chi-squared, etc)
plot data residuals
Notice that the help system in
XSPEC is started with
help, and works
similarly to the old good VMS help system.
Last Updated: 16 April 2010
