Main differences between SMART and SPICE

Disclaimer: This page aims at explaining the main differences between the IRS softwares SMART and SPICE. The focus is not on the comparison between the quality of extracted spectra, as users have different expectations. Here, the features that SMART can provide are emphasized.


SMART is an IDL package, enabling data extraction (tapered column, full slit, or optimal, extended source calibration, etc...) as well as a number of spectral and image tools. These includes detector image viewer, spectral analysis tool (line measurements, defringing, overplotting, average, spectra and image combination, etc...). Calls to the IDL programs IRSCLEAN and PAHFIT are also implemented to allow users easier extraction and analysis. Link to SMART.

SPICE is an extraction software (written in java) released by the Spitzer Science Center. It enables regular extractions as well as optimal extractions of high-resolution and low-resolution IRS data. Point-like sources or extended sources can be extracted. Data can be extracted as a single job or as a batch job. A simple spectrum viewer is also included. SPICE benefits from the same environment as Spot and Leopard. Link to SPICE.

Common elements

Both softwares use the same calibration files which are selected automatically based on the data image version. Both softwares run on a variety of platforms/OS and are scriptable, although to different extents.

The tapered column extraction of SMART is strictly equivalent to the regular extraction of SPICE. Spectra extracted either way are identical. Similarly, extended source extractions work the same way and give identical results. The main difference concerning the data extraction in both softwares is the implementation of optimal extraction.

Optimal extraction

SPICE currently enables optimal extraction of the 4 modules (SH, LH, SL, LL), while SMART enables optimal extraction of SL and LL.

The main difference between the 2 algorithms concerns the use of a super-sampled point spread function (PSF) in the case of SMART, while SPICE uses a PSF normally sampled (same as observed data). Since the SPICE PSF was determined at the reference positions (the 2 nods and the slit center), SPICE optimal extractions can result in wiggles if the source is not aligned with the reference positions. On the other hand, the SMART super-sampled PSF enables extractions anywhere along the slit. It also enables extraction of multiple sources, even blended.

SMART deals with the detector rows, contrary to SPICE which considers pseudo-rectangles (defined as the zones in the detector array with constant wavelength). For this reason, the wavelength array output by SMART varies as a function of the source position. This allows extracting the 2 nod spectra at slightly different wavelengths, which results in a better sampling of the spectra. In addition, considering rows instead of pseudo-rectangles results in a slightly larger spectral resolution (couple of percents).

Note that both SPICE and SMART optimal extractions are sensitive to the intra-pixel response function. SPICE because it deals with pseudo-rectangles and the response function is assumed to be uniform when pixels are split. SMART because the PSF is sampled at less than a pixel, which also results in some uncertainties.

Other unique features enabled by SMART: SMART optimal extraction remains possible for sources with significant offset in the dispersion direction. SMART enables the optimal extraction of point-like sources embedded in complex (extended) emission backgrounds. The algorithm can deal with multiple sources and a complex background. Finally, SMART source finder benefits from the super-sampled PSF to reach a position accuracy of better than a fiftieth of a pixel.

See the SMART optimal extraction page for more information.


  • SMART is organized around projects, which contain data files (spectra and images) and which can be saved. SPICE deals with extraction batch jobs which can be saved for later use.
  • SPICE enables extractions step by step making it possible to stop anytime for diagnostics.
  • SMART calls IRSCLEAN for an automatic cleaning of data images based on the BMASK and the campaign rogue mask.
  • Users in SMART can automatically combine individual exposures (DCEs).
  • SMART dataset manager allows users to combine images or make operations (e.g., subtraction, for the sky)
  • SMART includes IDEA, a spectrum viewer and analysis tool. IDEA enables line flux measurements, defringing, spectra combination, average, trimming, etc... Recently, a call to PAHFIT was implemented, enabling a quick look at the fit along with the output measurements (PAH, lines, dust, silicate).

This site is powered by FoswikiCopyright © 2002-2018 by the IRS Team.
Send feedback. We use Foswiki.