Spectrometer Open File Format

The Spectrometer Open File Format is a proposed file format that can describe a hyperspectral observation acquired with multiple spacecraft instruments including (but not limited to) mapping point spectrometers (e.g. Mars 2020 SHERLOC and PIXL, MSL ChemCam), multispectral imagers (e.g. Mars 2020 Mastcam- Z), and hyperspectral remote sensors (e.g. Cassini VIMS). Critically, a SOFF file is a supplemental description of a data set - describing existing data rather than repackaging or reformatting existing data. A common data file format will allow for the use of a single tool or visualization program to access instrument data from multiple instruments. SOFF is particularly designed to work with data in the PDS archives, although it is not limited to this. An especially powerful application of the adoption of SOFF would be the investigation of a single target observed from multiple instruments (as is the case on Mars 2020 with PIXL, SHERLOC, SuperCam, and Mastcam-Z) using a single visualization tool, which would not need instrument-specific readers to access each unique dataset - all while utilizing the existing data formats. A single tool could effectively inform target selection during mission operations by providing simpler access to all multi-spectral data products without requiring specialized knowledge of each instrument's data format.

The SOFF file is a manifest file which describes all components of an observation, for example, raw spectra, positioning information, calibration tables, context images, reduced products (e.g. elemental composition), and other relevant metadata (e.g. temperature values or source characteristics). It accomplishes this by pointing into existing data products and describing their format in common terms. The SOFF file organizes components of an observation by their dimensions, allowing a user to easily identify and group products according to a common independent variable. SOFF organizes hyperspectral products within an observation and improves the accessibility of data by presenting the user with a common data structure, independent of the original instrument data products.