Re: Fitting background-subtracted grating data in isis 1.4.9-4

On Wed, Jan 30, 2008 at 14:51 +1100, Duncan Galloway wrote:
> Hi,
>   I am trying to fit combined +/- order spectra once the background  
> has been subtracted off, and just can't work out how to do this in  
> isis. I've used add_grating_orders to produce the single pha (type I)  
> file from the pha2 file, and then tg_bkg to produce a (single)  
> background file from the output (this gives errors, but appears to  
> work). Then I load in my source file and assign the background:
> 
> isis> define_back (1,"sum_MEG_1_bkg.pha");
> 0
> isis>
> 
> But this does not appear to actually subtract the background 

Isis does not subtract the background from the data.  Instead,
the (scaled) background is added to the model.  The fit
statistic should be the same, since
   (data - background) - model = data - (model + background)

To plot the background-subtracted spectrum, one can retrieve
the numerical values using

  get_data_counts, get_back, get_model_counts

and generate a custom plot using hplot.  Mike Nowak has written
custom plotting routines that do this.  This plotting
functionality will eventually be incorporated into the isis
distribution.

> (does the "0" mean it's not actually doing it?) 

The define_back() function returns an integer status code to
indicate whether or not the function succeeded.  A return value
of zero indicates success (-1 indicates failure).

> Am I going about this completely the wrong way? Is there a
> simpler way, e.g. one that just uses the background columns in
> the pha2 file? any suggestions much appreciated! Cheers Duncan

Using combine_datasets to combine +/- order spectra has a
number of advantages over creating a summed pha file. For
example, combine_datasets allows each individual spectrum to be
treated with the appropriate (separate) response files so
there's no need to create summed response files. With
combine_datasets it's also easier to examine the effect of
leaving one or more spectra out of the sum.  For a more
detailed discussion, see

   http://space.mit.edu/cxc/isis/combine_datasets.ps

Thanks,
-John
----
You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
unsubscribe