Plotted quantities¶
Status: available.
What it is¶
The value plotted, shown as color on a raster or as the Y value on a scatter, can be:
Quantity |
Raster |
Scatter |
Meaning |
|---|---|---|---|
Amplitude, Phase, Real, Imaginary |
yes |
yes |
the complex visibility, averaged as Averaging describes |
Flag |
yes |
the fraction of each cell that is flagged |
|
yes |
yes |
how steady the phase is |
|
yes |
how far each sample is from the mean of its neighbours in time |
||
yes |
yes |
how unusual each sample is for its baseline |
|
U, V |
yes |
baseline coordinates, for UV coverage |
Phase is in degrees, from −180 to 180.
What it is for¶
Amplitude and phase show what the data are. The derived quantities show particular kinds of trouble:
Phase RMS and Coherence: phase instability (atmosphere, a bad antenna).
Amp V Diff and Phase Diff: short-lived changes (interference, a jump, one bad integration).
Z-Score: outliers, judged against each baseline’s own behaviour.
Flag: where flags already are.
The AIPS cookbook gives similar advice for TVFLG: amplitude finds problems lasting a scan or more, and the vector difference finds short excursions.
How it is calculated¶
Which samples count. Only unflagged samples enter any value. Every
unflagged sample counts equally: the data weights (WEIGHT,
WEIGHT_SPECTRUM) are not used in any average or statistic. AIPS and
CASA normally weight their averages, so on data whose weights vary,
visplot’s averages can differ from theirs.
Parallel hands. When a raster shows the parallel hands averaged, the visibility used is \(V = (V_{RR} + V_{LL})/2\) (or XX and YY). A sample is treated as flagged if either hand is flagged.
A raster cell. A raster cell covers a set \(S\) of unflagged samples: those that fall in it along the dimensions not shown, such as all selected channels on a Baseline × Time raster. With \(N\) samples in \(S\):
Quantity |
Vector averaging (default) |
Scalar averaging |
|---|---|---|
Amplitude |
\(\left\lvert \frac{1}{N}\sum_S V \right\rvert\) |
\(\frac{1}{N}\sum_S \lvert V\rvert\) |
Phase |
\(\arg \sum_S V\) |
\(\arg \sum_S V/\lvert V\rvert\) (each sample counts equally, whatever its amplitude) |
Real, Imaginary |
\(\frac{1}{N}\sum_S \mathrm{Re}\,V\), \(\frac{1}{N}\sum_S \mathrm{Im}\,V\) |
the same |
Phase is in degrees, in (−180°, 180°].
Flag is the fraction of the cell’s samples that are flagged, counted over all the selected correlations. Padding (time slots a baseline does not have) is not counted.
A cell with no unflagged samples is left blank.
A cell covering several baselines (any raster without a Baseline axis) is described under How it is calculated.
The derived quantities are described on their own pages: How it is calculated, How it is calculated and How it is calculated.
A scatter point is one sample, unless scatter averaging is on (see How it is calculated).
From Python¶
visplot(vis="obs.ms", raster_qty="AMP_VDIFF", scatter_y="PHASE")
Raster values for raster_qty: AMPLITUDE, PHASE, REAL,
IMAGINARY, PHASE_RMS, COHERENCE, AMP_VDIFF, PHASE_DIFF, FLAG,
Z_SCORE. Scatter values for scatter_y: AMPLITUDE, PHASE, REAL,
IMAGINARY, U, V, PHASE_RMS, COHERENCE, Z_SCORE.
Limits¶
Weights, and calibration quantities (gains, delays, Tsys), cannot be plotted yet.
There is no amplitude RMS quantity (see Phase stability: Phase RMS and Coherence).
Coming from plotms or AIPS? See Coming from plotms and Coming from AIPS.