next up previous contents
Next: Cookbook Up: Data Analysis Previous: Data Analysis   Contents


The first task when starting a SPI data analysis is to identify and collect the input data. There are two main cases, (1) guest observers receiving their private data, and (2) users of the public data, browsing and retrieving data from the INTEGRAL archive. Basically, a list of pointings have to be collected into an ASCII file. This ASCII file can then be fed into og_create to produce what is called an observation group (OG). The OG includes links to all input (and later output) data, and all software read/write data through this group.

Once you have created an OG, an image reconstruction or a spectral extraction can be achieved by running a single script ``spi_science_analysis''. This script analyzes the data going through a number of different steps briefly described in Table [*]. These steps correspond to so-called ``Levels'' identified with 3-5 capital letter acronyms. Their meaning is common to all INTEGRAL instruments. An important difference of SPI with respect to the other instruments is however that (almost) always the complete set of pointings are analyzed together simultaneously. Except for the GRB analysis, single pointings are never analyzed independently.

Table: Overview of the successive steps of the spi_science_analysis script
Level Program Short description of the step
CAT_I cat_extract Extracts from the INTEGRAL reference catalogue information (position and flux) about known sources lying in the field-of-view.
POIN spi_obs_point Extracts the start/stop times and the SPI attitude for all pointings included in the observation, and stores the results in the pointing.fits file. On request, this program can access a list of bad pointings (usually of order of 1-2% of the total) and exclude them for the current analysis
BIN_I spibounds Defines the energy bins creating the energy_boundaries.fits file.
  spi_obs_hist On-the-fly energy correction, event binning, and derivation of the dead times. Outputs three files (1) gti.fits containing ``always good'' GTI with on-times, (2) dead_time.fits containing dead time ratios and livetimes, (3) evts_det_spec.fits containing the binned events formatted as spectra.
  spi_phase_hist (Optional) event binning in phase intervals computed from provided ephemeris and taking into account possible orbital motion. Outputs one file per phase bin named evts_det_spec_i.fits where i is the phase bin number.
BKG_I spi_obs_back Computes background models from different indicators. Two main recommended options are (1) deriving the background from pre-defined flat-field files, and (2) using the Germanium saturated events as a background tracer. Outputs are stored into back_model.fits.
IMA, or SPE, or LCR spiros Carries out image reconstruction (IMA), spectral extraction (SPE) or timing analysis (LCR) depending on the chosen ``mode''. This is the most difficult and critical step. If the phase binning as been selected, all phase bins are processed independently in turn.
  spirmf Invoked in spectral extraction only, creates the appropriate RMF response matrix required for XSPEC spectral fitting.

Since the release of OSA 8, another script is available for the spectral extraction and the contemporary fit of diffuse emission: spimodfit_analysis. This script performs all the steps of spi_science_analysis before running spiros. Instead of this program, the script calls the program spimodfit which allows for a different approach in the simultaneous fit of background, point sources, and diffuse emission using the SPI data. This will be described in Sect. [*] of the cookbook and with more detail in Sect. [*]. Finally a dedicated script spi_grb_analysis is available for the analysis of GRBs, and its related on-line documentation can be found at

next up previous contents
Next: Cookbook Up: Data Analysis Previous: Data Analysis   Contents