MIDI pipeline is still under development. This version of the pipeline processed dispersed fringes taken in HIGH_SENS mode
and produced dispersed visibility measurements.
- CALIBRATION
DATA
- Technical calibration data:
There are no recipes to reduce the technical calibration data.
- Detector dark frames are taken as requested in the calibration
plan and distributed to the user.
- Wavelength scale has been calibrated by the MIDI group in december
2003 and the calibration is included in the "IMAGE_DETECTOR"
table included with all raw MIDI data
- Astronomical Calibration data:
The recipes used to produce astronomical calibration data are :
- "reduce fringe" which calculate an uncalibrated Visibility
using the TRACK and its associated PHOTOMETRY raw files.
The raw spectruum is collapsed using a weighted mask to yield to a spectruum one one raw.
In this version of the
pipeline the mask applied to the data is a fixed Mask corresponding
to the instrument mode. Each scan is individuadly processed and
can be rejected if no fringes are found. The raw visibility is
calibrated using the photometric channel.
- "Transfer_Function" which look at the VLTI catalog
to find the estimated angular diameter of the astronomical calibrator.
An instantaneous instrumental transfer function is calculated.
The final products are : a FITS product file whose primary headers
contains a set of QC Paramaters
, plots of the dispersed Visibilities, and an ascii file containing the value of the instantaneous Transfer
Functions measured on the VLTI calibrators.
- SCIENCE
DATA
The recipes used to produce science data are :
- "reduce fringe" which calculate an uncalibrated Visibility
using the TRACK and its associated PHOTOMETRY raw files.
The raw spectruum is averaged in one direction using a weighted mask (pipeline
recipe called "aperture_mask"). In this version of the
pipeline the mask applied to the data is a fixed Mask corresponding
to the instrument mode. Each scan is individuadly processed and
can be rejected if no fringes are found. The raw visibility is
calibrated using the photometric channel.
- "Calibrate_Visibility" which takes the calibration
product associated to the science product (see association
rules) and uses the Instrumental Transfer function calculated
on this astronomical calibrator to produce a Calibrated Visibility.
For the scientific data the final product is a FITS product file
whose primary headers contains a set of QC
Paramaters.
This table gives a summary of the pipeline recipes and their status
raw data category
|
pipeline recipe
|
product
|
comment
|
status
|
SCIENCE - CALIB
|
aperture_mask
|
intermediate
|
define the mask to be applied to the fringe
files
|
tested
|
SCIENCE - CALIB
|
reduce_fringe
|
final : FITS file containing uncalibrated Visibility point
|
calculate the uncalibrated visibility
|
tested
|
CALIB
|
Transfer_Function
|
final : ascci file + FITS file updated with
TF
|
calculate the Transfer Function of the instrument
for this observation
|
to be finalized
|
SCIENCE
|
Calibrate_Visibility
|
final : FITS file updated with calibrated
Visibility point
|
calculate the calibrated visibility for
this observation using a TF calculated on a calibrator
|
to be finalized
|
The results
should be used only for Quality Control purposes.
The Pipeline processed dispersed fringes in dispersed mode. The Visibility written in the products is a Vis^2.
The error on the raw visibility produced by the pipeline depends mainly
on the stability of the photometric data. For object fainter than 15 Jy,
the error on the visibility is dominated by the instability of the
photometric beam. For objects brighter, the accuracy in the photometric
calibration done by the pipeline limits the error in the raw
Visibility to about 15%.
In some cases, the Visibility calculated in white light is greater than unity in the pipeline product as well as with the off line MIDI Data reduction packages. It could be related to a chopping problem during the photometry. Check the ESO HIERARCH ISS CHOP THROW keyword, which, when set to zero, indicates a problem with chopping. (This does not necessarily mean that the template contained a value of zero for the chopping throw.) This reduces the flux in the chopped spectrum and consequently leads to a normalization of the fringe spectrum with a flux too small, causing larger than real visibilities.
|