How to reduce data taken in Sci_Phot mode

MIA+EWS now has some facilities to reduce MIDI SCI_PHOT data that have been initially tested. This page describes at an elementary level the available procedures. Note that this is still rather experimental, there is no guarantee, explicit or implicit, that anything works, but let us know if you discover bugs.

Cross-Coupling Coefficients

It is necessary to calculate the cross-coupling (kappa) coefficients between the I1+I2 channels and the PA+PB channels. We don't know whether these are stable for hours, days, years... (if shorter than hours there is no point in doing SCI_PHOT mode). To calculate them you need standard chopped photometry data (i.e. with AOPEN and BOPEN shutter settings, taken with the NRTS mode DEFAULT_CHOP) on a STRONG CALIBRATOR. You can in theory combine several (many) such observations over a period of a day if you think you know what you are doing:

    midiCrossCoeff, base, files

• base
  is a character string containing the beginning of the output file name, include directories, e.g. base='/home/user/testdata/cc', in which case the output file is /home/user/testdata/cc.crossCoeff.fits.

• files
  is an array of two character strings, describing the AOPEN and BOPEN data. If there is more than one file in each group (either because MIDI broke up the output into multiple files, or because you want to combine several observations), then the files should be concantenated into one string, but separated by blanks. This is the form usually delivered by midiGui(s).

Reducing Interferometric Data

Assuming the calculation of the kappa coefficients worked you now have four (4!) different possible modes to reduce your science target (and calibrator) data:

A. No standard photometric data

If you have NO standard AOPEN/BOPEN (DEFAULT_CHOP) data on your target, then the photometric data will be taken from the PA+PB channels of the interferometric data. This has the advantage that it saved you a lot of time at the telescope, but the disadvantages:

• The image quality of the PA/PB spectra (especially in GRISM mode) is lousy at the corners of the pictures.
• There are small wavelength shifts between the PA/PB channels and the I1/I2 channels that I have not yet corrected.

Anyways, to reduce data in this format with EWS type:

    midiSPipe, base, file, cross=crossFile

file is a single (one!) string, describing the ABOPEN data taken in interferometric mode (i.e. your DISPERSED_FRINGE_TRACK data).

You can also add the usual fudging parameters like smooth, gsmooth, dSky, dAve...

The output is the usual array of files base.corr.fits, base.photometry.fits... plus some new intermediate result files like base.Aphotometry.fits, base.ABphotometry.fits

What the program does is to resample the PA+PB data onto a grid so that it looks like the I1+I2 data, and then run the standard masked photometry reduction on this data.

The messages on the screen will look different than older versions of midiPipe; I have tried to print more diagnostics and standardize the data stream.

To reduce the same kind of data with MIA type:

    x = obj_new("xmidispvisi",file,cross=crossFile)

file and crossFile are the same parameters as for EWS. In fact, MIA is using the same programs as EWS to convert the PA+PB data to I1+I2 data, and run the standard masked photometry reduction on this data.

B. Sci_Phot Fringe Track plus Standard Photometric Data

You have some standard AOPEN/BOPEN (DEFAULT_CHOP) data on your target, that you would like to use to define the details of the photometry data; but you want to use the PA+PB data taken simultaneously with the interferometry data to monitor atmospheric changes and overlap changes. What we do is reduce the DEFAULT_CHOP photometry as if this were HIGH_SENS reduction (see section D. below). Then we find the coefficients C and D so that this spectrum * (C+Dx) is as close as possible (in the least square sense) to the PA+PB photometry taken simultaneously with the interferometry. We use this "rescaled" photometry in all future reduction.

The calling sequence for EWS is the same as above:

    midiSPipe, base, files, cross=crossFile

With MIA you can use xmdv:

    x = xmdv(files, cross=crossFile)

C. Independent Photometric Data

As in (B) but you have some really fantastic photometric data (AOPEN/BOPEN) on your source that you reduced previously, with midiPhotoPipe (see below), and you want to use this as the "detailed" photometric signal for a whole set of interferometric runs. Call:

    midiSPipe, base, file, cross=crossFile, photoFile='fantastic.photometry.fits'

To reduce the fantastic photometry first, use:

    midiPhotoPipe, base, files [,dSky = dSky]

Suppose you want to average together five (5!) sets of photometry data of a source (all in the same GRISM/PRISM mode, and all the same source). You have selected them using midiGuiS:

    f = midiguis()

    nf = n_elements(f)
    fa = f[0]
    for i=2,nf-2,2 do fa=fa+' '+f[i]
    fb = f[1]
    for i=3,nf-1,2 do fb=fb+' '+f[i]
    fab = [fa,fb]

    midiPhotoPipe, base, fab

With MIA, method B and C are (almost) equivalent, since the de-chopped photometric files (the equivalent to EWS' "fantastic.photometry.fits") are stored in the PHOTOMETRY-directory. To use five sets of photometry data like in the example above, create the strings fa and fb, and put them into an array with the name of the interferometric data:

    files = [ fringeFile, fa, fb ]
    x = xmdv(files, cross=crossFile)

D. Reduce Sci_Phot data in High_Sens mode

You just want to pretend that this is HIGH_SENS data, and ignore the PA/PB channels:

    midiPipe, base, files

MIA can be used in the standard way, the only difference to real Sci_Phot-mode is the absence of crossFile:

    x = xmdv(three_files)