Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.eso.org/~chummel/oyster/manual/node165.html
Дата изменения: Tue Apr 28 19:12:50 2015 Дата индексирования: Sun Apr 10 20:38:41 2016 Кодировка: Поисковые слова: п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п п р п р п р п р п р п р п |
The CALIBRATEVISIBILITYCALIBRATE button is for the calibration of visibilities (squared amplitudes, triple amplitudes, and closure phases). You will be presented a widget in which to select the calibrator stars and system visibility indicators. As for the calibrator stars, there are three options: to select all (an unusual case), select according to a flag in the startable (derived from information contained in file diameter.bsc), or to select manually (recommended).
For each selected indicator, a list of base function will be displayed and you will select a minimum number of functions necessary to represent the dependence of the visibility on this indicator. The final fit will be multi-dimensional, with all functions fitted to the calibrator data simultaneously. (Exception: the selection of a smoothing function, see below!) The mathematical procedure used is the Singular-Value-Decomposition (SVD). In case of a design matrix singularity, SVD will edit the small eigenvalues and proceed to calculate a solution for the calibration.
The smoothing functions S_XX (the index gives the width in minutes) are available for hour angle and time calibrations. Selection of these disable any other selection since they cannot be used in the multi-dimensional fit. This is brought to the attention of the user by making all other base function widgets insensitive while a smoothing function has been selected.
Which indicators and functions are to be used in the calibration depends on atmospheric and instrumental circumstances and has not been fully investigated for the NPOI yet. Therefore, even though a wide array of indicators is available, only one or two might typically be used. Ideally, the visibility amplitude should depend only on the delay jitter, but instrumental effects might require an additional time dependent calibration, either depending on time itself, or on hour angle if the delay line position or the siderostat mirror pointing has played a role in the visibility degradation.
Before the actual calculation of a solution it is required to select the data which is to be used for the computation of the calibration coefficients and is to be calibrated. Clicking on CALIBRATEVISIBILITYPLOT will display the plot widget. Please note that the star selection in this widget will now indicate the stars for which the data is to be calibrated. It should also include the selected calibrator stars! Since several calibrations can be applied sequentially to the same data, appropriate selections are required here if calibrations differ from star to star. Also note that the scan selection is used just as in the case of plotting scan data. The calibration is only applied to valid visibilities. By selecting the LOOP option, selected channels and baseline will be calibrated independently rather than combined in a simultaneous fit. This options was implemented to speed up calibration of an entire spectrometer.
Remember to work on all spectrometers and baselines. In the case of NPOI 6-way data, combine only scans of the same sub-array configuration. Unless the photometric calibration is very good, sub-arrays will differ in their system visibility.
Look at the normalized and calibrated visibilities to make sure they were not calibrated before if you want to start over with a new calibration. This is because, as mentioned above, more than one calibration can be applied sequentially, and calibrated data is part of the data saved in the output.
Click on CALIBRATE in the calibration widget to compute a solution based on the calibrated estimated visibilities (amplitudes or closure phases, depending on your selection) of the calibrator stars. (The calibrated visibilities are initialized as a copy of the uncalibrated visibility.) The results will be entered in the calibrations table (strictly speaking an array of calibration entry structures), and automatically applied to the data. The entire calibration table can be stored to disk using ENTRIESSAVE, and loaded from disk using ENTRIESLOAD. A list of all calibration entries will be printed to screen by ENTRIESLIST.
Here we summarize again the calibration widget functions:
UNDOUNDO will remove a specific calibration from the data as well as the corresponding entries in the calibration table. You will be presented with a list of calibrations for the selected variable, from which to choose the calibration to be removed. The data selection is provided by the information stored in the calibration table. (To remove a calibration only from the table, use the ENTRIESEDIT function.) UNDOUNCAL will remove all calibrations for the selected data and all stars/scans. UNDORESET resets the calibration table (for amplitudes or phases, depending on your selection) and remove all calibrations from all data.