It is useful to be able to make a rough and ready reduction of your observations at the telescope, if only because you will want to check -- before starting your set of integrations -- that you are observing the right object! From your finding chart, the telescope operator will have acquired your field using the acquisition bundle of INTEGRAL. However, you may wish to check that your object of interest is positioned appropriately in your field of view by looking at the reconstructed image of your 3D spectra. This could also be used to check for any drift over time of your object off of the field of view. These INTEGRAL data reduction procedures are run in IRAF, you call them up by typing
cl>integral
We recommend that you copy your data to another directory before you manipulate them.

For a quick look of data during your run you can do the following (check out also the data reduction quick guide for fuller details on how to run the individual INTEGRAL procedures):

1. You need a lamp flat-field image to identify and trace the fibres, and thus extract the spectra: one for each fibre bundle you will use (and new ones if you change the slit translation). You should have created these flats before starting observing. The wavelength range is unimportant for tracing purposes, however set toward the red end of the spectrum in order to get enough counts in the blue (with some gratings, it may in fact prove impossible to get enough counts in the blue end). To create these flats:

Let say you are using fibre bundle SB1 (called STD1 below). In lpvs1 turn on the white lamp and in your IRAF xterm take a flat exposure:

                  ICL> ila w on
                  ICL> isp std1 cal  
                  cl> flat integral 2 "flat STD1 template" 

Give this flat-field image a memorable name and save it. Remember to turn off the light when you have finished.

2. Use the int_apall task to define and trace the apertures from the flat-field image, specifying a blank (ie. no) reference flat. See the data reduction quick guide for how to do this (and we note here that to work with your data, you will probably have to change the DISPAXIS header paramter to 1 as it does not automatically change when you rotate the CCD from the WYFFOS position). With the 2-CCD setup, and the resulting gap between, one either has to trace the spectra (and later extract) very roughly, or treat the 2 CCDs individually.

3. Take an exposure of your object, eg.

                     ICL> isp sdt1 obs
                     ICL> run integral 200 "test object"

4. Use int_apall to extract the spectra from this object frame. Specify for the reference flat that which you previously just traced, and no for every other parameter with yes/no entries, except for recentre and extract. Enter "no" to each question the task asks. The output is a file with the same name as your object frame; but with .ms at the end (e.g. r294021.fits -> r294021.ms.fits). This file is an MxN pixel image, where N is the number of fibres in the bundle and M the number of pixels in the dispersion direction.

5a. Use the imarec task to create a (basic) reconstructed image from your spectra. Most of the inputs are self explanatory, and many you can leave alone. The map pixel scale is the number of arcsec per pixel of the reconstructed image. Chose a value less than the actual arcsec-per-pixel of the fibres of your bundle. You can also specify the pixel range (in the dispersion direction) to use to create the map (eg. to encompass an interesting emission line). ivalue set how the counts in each spectrum are added up to create the fluxes in each pixel of the map. In broken you should enter "no", not yes. The output of this task is an image which you can display normally.

5b. OR use the Euro3D visualisation tool to image the data. This will allow you to create a spectral-image of your data, and to inspect the spectra of each fibre (or "spaxel" in Euro3D parlance). To run this, you have to say how when it has been installed at the ING. In the data reduction manual are quick instructions for using this visualisation tool, for a more detailed manual, see the manualput when got it and got a directory to put it in.

6. For extracting and then calibrating arc exposures, we recommend that to avoid the gap between the two CCDs messing it all up, either treat each CCD individually, or have traced the flatfield spectra (and thus extract the arc spectra) very roughly (as was discussed above). Experiment with this; for blue spectra I found that it was not possible to wavelength calibrate the red and blue CCDs together, they had to be treated separately. However, you may not need to calibrate the data that exactly (it can be done surprisinglly accurately by eye).