This application converts a SPECX map file into a simple data cube formatted as a standard NDF. It works on map files in Version 4.2 or later of the SPECX format. It can optionally write a schematic of the map grid to a text file.

In addition, it will also convert an HDS container file containing an array of one-dimensional NDFs holding SPECX spectra into a similar container file holding individual, scalar NDFs each holding a single spectrum from the supplied array.

In both cases, WCS components are added to the output NDFs describing the spectral and spatial axes.

A VARIANCE component is added to the output NDF that has a constant value derived from the Tsys value, integration time, and channel spacing in the input.

AXIS = _LOGICAL (Read)

GRIDFILE = LITERAL (Read)

IN = NDF (Read)

LATITUDE = LITERAL (Read)

LONGITUDE = LITERAL (Read)

OUT = NDF (Write)

SYSTEM = LITERAL (Read)

TELESCOPE = LITERAL (Read)

specx2ndf specx_map specx_cube

This example generates an NDF data cube called specx_cube (in file specx_cube.sdf) from the NDF SPECX map called specx_map (in file specx_map.sdf). A text file containing a schematic of the map grid will not be produced.

specx2ndf specx_map specx_cube gridfile=map.grid

This example generates an NDF data cube called specx_cube (in file specx_cube.sdf) from the NDF SPECX map called specx_map (in file specx_map.sdf). A text file containing a schematic of the map grid will be written to file map.grid.

SPECX map files are written by the SPECX package (see SUN/17) for reducing spectra observed with heterodyne receivers operating in the mm and sub-mm wavelength range of the electromagnetic spectrum. SPECX is usually used to process observations obtained with the James Clerk Maxwell Telescope (JCMT) in Hawaii.

A SPECX map file comprises a regular ‘rectangular’ two-dimensional grid of map positions on the sky, with spectra observed at the grid points. However, a spectrum is not necessarily available at every grid position; at some positions a spectrum is not observed in order to save observing time. For example, for a grid centred on a typical, roughly circular, object spectra may be omitted for the positions at the corners of the grid. SPECX map files are standard Starlink NDF HDS structures. The principal array of the NDF is a two-dimensional array of the grid positions. The value of each element is either a pointer to the spectrum observed there (in practice the number of the spectrum in the array where they are stored) or a value indicating that a spectrum was not observed at this point. In effect the SPECX map structure is an implementation of a sparse array.

SPECX2NDF expands a SPECX map file into a simple three-dimensional data cube, again formatted as a standard NDF, in which the first and second pixel axes corresponds to the spatial axes and the third axes correspond to the spectral axis. The advantage of this approach is that the converted file can be examined with standard applications, such as those in Kappa (see SUN/95) and easily imported into visualisation packages, such as Data Explorer (DX, see SUN/203 and SC/2). When the output data cube is created the columns corresponding to the positions on the sky grid where spectra were not observed are filled with ‘bad’ values (sometimes called ‘magic’ or ‘null’ values), to indicate that valid data are not available at these positions. The standard Starlink bad value is used. Because of the presence of these bad values the expanded cube is usually larger than the original map file.

The created NDF cube has a WCS component in which Axes 1 and 2 are RA and DEC, and Axis 3 is frequency in units of GHz. The nature of these axes can be changed if necessary by subsequent use of the WCSATTRIB application within the Kappa package. For compatibility with older applications, AXIS structures may also be added to the output cube (see Parameter AXIS). Axes 1 and 2 are offsets from the central position of the map, with units of seconds of arc, and Axis 3 is frequency offset in GHz relative to the central frequency. The pixel origin is placed at the source position on Axes 1 and 2, and the central frequency on Axis 3.

SPECX2NDF reads map files in Version 4.2 or later of the SPECX data format. If it is given a map file in an earlier version of the data format it will terminate with an error message. Note, however, that SPECX itself can read map files in earlier versions of the SPECX format and convert them to Version 4.2.

SPECX2NDF has an optional facility to write a crude schematic of the grid of points observed on the sky to an ASCII text file suitable for printing or viewing on a terminal screen. This schematic can be useful in interpreting displays of the data cube. It shows the positions on the grid where spectra were observed. Each spectrum is numbered within the SPECX map structure and the first nine are shown using the digits one to nine. The remaining spectra are shown using an asterisk ("*"). You specify the name of the file to which the schematic is written. Figure 1 shows an example of a schematic.

SPECX2NDF copies all the auxiliary information present in the original map file to the output data cube. However, the arrays holding the original spectra are not copied in order to save disk space.

In addition to converting SPECX map files, this application can also convert HDS files which hold an array of one-dimensional NDF structures, each being a single spectrum extracted by SPECX. Since arrays of NDFs are not easily accessed, this application extracts each NDF from the array and creates a new scalar NDF holding the same data within the output container file. The name of the new NDF is SPECTRUMn where n is its index within the original array of NDFs. Each new scalar NDF is actually three-dimensional and has the format described above for an output cube (i.e. Axes 1 and 2 are RA and DEC, and Axis 3 is frequency). However, pixel Axes 1 and 2 span only a single pixel (the size of this single spatial pixel is assumed to be half the size of the resolution of the JCMT at the central frequency). Inclusion of three-dimensional WCS information allows the individual spectra to be aligned on the sky (for instance using the Kappa WCSALIGN task).