Basic data reduction and observing instructions

The following INTEGRAL IRAF commands are likely to be used at the telescope:

int_apall, imarec, int_fwhm1 or int_focus, int_slit, int_center and in addition there is the Euro3D visualisation tool

Unless otherwise stated, the parameters should be left at the default values. We similarly do not detail the obvious inputs.

int_apall
This is to identify, trace, and extract the fibres (apertures) from the raw image. The defaults set for this task are to allow you to extract the data from the combined-CCD setup, ie. with the window set so that when you display your image, both CCDs come up on the display tool (eg. DS9). Because of the gap of 105 pixels, it is difficult to trace the apertures on both CCDs, and for your proper data reduction we recommend you treat the two CCDs separately. However, for use at the telescope, we have set the parameters t_nsum and t_step to 60, and t_nlost to 10, so the tracing will miss the gap (if it does not, fiddle with these numbers until it does). It will suffer by not following all the real wiggles of the apertures on the CCD. In addition, the fading of the spectra at the very red end will result in some sharp jumps in the traces. For the purpose of simply looking at your spectral-images this is not important, as you can ignore the first 100 pixels or so.

For bundle STD3 we have allowed for two setups, one without any binning (option "STD3" in int_apall) and one with binning by a factor 2 in the spatial direction (option "STD3b" in int_apall). The binning factor in the spectral direction is unimportant for this task, so it will work with or without binning.

If you are using int_apall to identify and trace fibres, do it with a dome (or sky) flat. The references should be left blank. The bundles are specified as "STD1/STD2/STD3/STD3b/COR". You do want to run interactively, edit, find, trace and recenter, and when asked, you do want to save to the database. The trace fitting order should be about 10.
When you run this program, it will open an interactive window that shows the slice of the centre of your image, with the fibre numbers written above. For STD1, STD3, STD3b and COR the fibres are counted from down to up (equals left to right), and for STD2 from up to down. It is your job to make sure that all fibres have been identified and assigned the correct number. First check the numbers given to the broken fibres (see Figure 1 here). If the broken fibres have been skipped, you will need to mark them, if the numbers assigned are not correct, the a fibre down the line has probably been skipped, although occasionally an extra fibre is mistakenly assigned at the very end or beginning. A skipped fibre is usually because it is just slightly mis-centred with respect to its neighbours; to correct the sequencing simply delete the offending aperture and re-identify it exactly half-way between the two adjacent. Then re-order the whole set. It may be necessary to play with the task parameters in order to force the task to accept your identifications; with STD1 this can be a problem as the fibres are unevenly spaced on the CCD.
int_apall is based on the IRAF task apall, and as with this task, typing "?" in the window will give you a list of the commands. Those you are most likely to use are:

shift-x to zoom on the cursor position
w-x, w-r, w-l, w-a to zoom and then be able to scroll right, left, or show the whole window
:parameter value to change some parameters, such as the llimit, ulimit, maxsep or width
d, n, o to delete, mark a new, and re-order apertures.

If the routine cannot find the apertures, it is possible is it looking in the wrong direction. The DISPAXIS header parameter of your files should be "1" (edit with ehead after copying the file to a test directory in which you have permission to create new files, which you will not have in the directory the data are sent to from the CCD).

Once all the fibres are identified, they are then traced. It is worth occasionally typing "yes" to view these traces, especially the central ones where the slope of the aperture/spectra on the CCD changes direction.

If you are using the routine to extract spectra from an astronomical image, then you do not want to run interactively, edit, find, or trace, but you do want to extract and recenter them. The reference image should be your aperture calibration flat-field, on which you have previously identified and traced the fibres. You should do this all in the same test directory you traced the flatfield from.

imarec
This is to create a map from your extracted spectra. At the telescope there is no need to do the fibre response level correction (fibre flat-fielding using the sky flat-field image) except for the coronographic fibre bundle. The input image should be the .ms image created by int_apall. The map pixel scale is the number of arcsec to be in each map pixel; we suggest values of 1/4 or so of the actual fibre sizes. You should answer "no" to the question about the broken fibres, even if the answer should be "yes". the created map can then be displayed on your image display tool.

int_fwhm1
This is to measure the fwhm of selected positions on an arc exposure, where arc lines lie. The positions are read in from an input file (in the format -- X Y -- with newline breaks for each set and no blank line at the end). The output to the xterm is the measured mean fwhm for all lines, and the dispersion, and a series of plots showing the Gaussian fits to the arc lines (the order of these plots is to go from left to right on the first line of plots, left to right on the second line, etc., following your X Y positions in the file, going down the list). The dispaxis should be set to 1 for vertical, not for horizontal as is currently stated in the parameters, and v.v.

int_slit
This is similar to int_apall, however it will take two slices of the image (which you specify: eg at Y positions 100 and 950) and show you the cut and allow you to view it in detail. This is to check that the apertures are all present on the CCD.

int_focus
This is for the focus testing using the left and right Hartmann shutters method. The dispersion axis parameter may be the same way around as for int_fwhm1; however it currently crashes DS9 when we try it, and so we do not recommend using routine. You will have to do this focus test by hand. The idea to to look for the difference between the fwhm for the left and right Hartmann shutter images; when close to 0, you have a focused spectrograph.

Visualisation tool
In addition, there is the Euro3D visualisation tool you can use to inspect the data. For this you must either have the data in .ms format (int_apall extracted spectra) which has been wavelength calibrated (preferable) or not (perfectly acceptable). A quick quide to using the vistool is:

1) First convert the data into Euro3D format with the command:
myint2e -if [input file name] -of [output file name] -bundle [1,2,3] -type 1 -pf [SB/1,2,3/.data]
we suggest you call your new file "[whatever].e3d". "-bundle" is for which bundle you are using, SB1,2 or 3. "-type 1" will exclude the sky fibres, which is a very good idea for visualising the data as otherwise the field of view will be too large. "-pf" is the position file for your bundle, where the filenames are "SB1[1,2].data".

2) Next, call for the Euro3D setup with equiv of e3dset and start the visualisation tool with blah

3) Load the data into the vistool with File and then Load. Load your ".e3d" file from the GUI brought up. You will probably need to adjust the max and min levels to see the spectrum nicely on the main window.

4) Open the spectral and spaxel inspectors by selecting Spaxels and Open, Spectra and Open.

5) Select the wavelength/pixel range you wish to make a map in, in the main (first) window, with two left clicks of the mouse, and send to the spaxel inspector with a right click. This map/image can be adjusted for max and min levels, as well as brightness and contrast. The image you will see will be of your data summed and averaged over the wavelength range you have selected.

6) To inspect the spectrum of any particular "spaxel" (a spaxel being the individual elements you can see on the image in the spaxel inspector window), run over them with the central mouse button depressed, or select a single spaxel with the left mouse and send to the spectral inspector window with the right mouse button (remembering to Select:Clear that selected spaxel if you then want to look at another one). Adjust the display scales to see the spectrum properly. To zoom in on a region, again first select a range with the left mouse button.

There is a lot more you can do with this tool, but we refer you to the manual). for further instructions.

Although we have assumed you are familiar with IRAF, a few hints about the wavelength calibration may prove useful. It will be necessary to carry out the following sequence of commands:

setjd with "utobs" for the time, on all extracted (.ms format) images.
refspectra (from noao.imred.specred) the astronomical images, to identify which calibrated arc image to use. This sometimes does not work, in which case one can manually edit the headers with
      hedit file.ms REFSPECTRA arc.ms (the wavelength calibrated one) and
      hedit file.ms REFSPEC1 arc.ms (the not wavelength calibrated arc image)
dispcor the spectra.