This is a complete list of all the parameters used by echomop tasks.
Note that in most circumstances there will only be two reference frames. One taken
prior to the object frame, and one subsequently. If you do not wish to provide an arc
frame at all then you may reply NONE is response to this prompt or, alternatively set
the parameter TUNE_NOARC=YES. If no arc is available then you must provide your own
sets of wavelength feature positions, or wavelength scales, if wavelength calibration is
required.
The default is echomop_output.txt.
AUTO=YES then echomop will use all the options available to automate the processes it
invokes. Currently the following parameters will be affected:
NO.
NO.
NO.
NO.
NO.
NO.YES then the package will try to identify the arc line features
automatically. For orders with sufficient features this will usually be successful (
12
lines), and a useful strategy is probably to do an automatic run first, and then manually examine
the fitted features to verify their suitability.
The default value is zero which is ignored, in case of a mistake.
Orders are disabled by writing the ‘bad value’ quantity into the first coefficient of the trace polynomial. This ensures that they also appear disabled to other tasks (almost all of which will need the trace polynomials).
To re-enable an order, re-fit a polynomial to its traced path, using ech_fitord/echmenu option 3.
Initial suggested value: 0.0.
The recognised functions are POLY (the default), MEDIAN and SPLINE.
Note that the required maximum number of coefficients is different in each case. For POLY the
maximum number of coefficients corresponds to the number of degrees plus one, so for a cubic
use coeffs=4. For a SPLINE the maximum number of coefficients corresponds to 2x(knots+7), so
for 3 knots use coeffs=20. If you are using SPLINE fits then you should ensure that the
value of TUNE_MAXPOLY is set accordingly, as well as using an appropriate value for
BLZ_NPOLY.
When MEDIAN is selected local medians of X-extent BLZ_NPOLY-pixels are used as the blaze
function.
Initial suggested value: POLY.
YES if the blaze fitting is to be done interactively. Automatic fitting is performed
otherwise.
Initial suggested value: NO.
The type of function to be used is set using the parameter BLZFIT (default is POLY for
polynomial).
When BLZFIT=POLY then the number of coefficients should be set to degree of polynomial +
1.
When BLZFIT=SPLINE the number of coefficients should be set to 2 x (number of knots +
7).
When BLZFIT=MEDIAN the number of coefficients should be set to the number of pixels over
which a local median value of the blaze function is to be found.
For example, setting BLZ_NPOLY to 4 would permit the fitting of polynomials of the form:
constant + ax + bx squared + cx cubed
Initial suggested value: 7.
Pixels may be either replaced with an interpolated value, or by a flag value.
For example, the paths of the orders as fitted with polynomials will be overlaid on a part of the traced image.
Initial suggested value: NO.
The database is built from a .ARC file using the task ech_ftrdb.
The database should have the same file name as the .ARC file used for its creation. The default is
to use $ARCDIRS/THAR. Note the use of the $ARCDIRS environment variable to specify the search
path for arc databases. This should normally be used as it searches in the normal fiGARO order.
If you wish to specify a private database then the full pathname may be provided, e.g.,
/mydisk/mydir/sub/ANARC.
If no arc database exists for the .ARC list you want to use then use the ech_ftrdb task to create a
copy. You might like to check to see if anyone local already has the database you
need.
Initial suggested value: ’$ARCDIRS/THAR’.
This file will normally be created by the first echomop task run. Any valid datafile name can be used.
By default the file is created in the current directory.
This may be one of the following:
O — Optimal (default).
S — Simple.
P — Profile weighted.
The package uses this frame to calculate balance factors on a per-pixel basis. If a balance frame is already available, then the frame should be supplied instead of a flat field, and the parameter TUNE_PREBAL should be set.
If you do not wish to supply a flat field at all, then you may either reply NONE to this prompt
(must be uppercase), or set the parameter TUNE_NOFLAT=YES.
If no flat-field frame is supplied then the balance factors will be set to unity for all pixels.
The available functions are MEAN (the default), POLY, SPLINE, MEDIAN, SLOPE and SMOOTH.
All modes except POLY and SPLINE use local neighbourhoods delineated using a sample size (in
pixels) defined by TUNE_FFLSMP.
Note that the required maximum number of coefficients for POLY and SPLINE are different. For
POLY, the maximum number of coefficients corresponds to the number of degrees plus one, so
for a cubic use coeffs=4. For a SPLINE the maximum number of coefficients corresponds to
2x(knots+7), so for 3 knots use coeffs=20. If you are using SPLINE fits then you should ensure
that the value of TUNE_MAXPOLY is set accordingly, as well as using an appropriate value for
TUNE_FFNXPLY and/or TUNE_FFNYPLY.
Initial suggested value: MEAN.
The trace frame is checked, and the other frames are assumed to have identical bad row and/or column features.
For example, ps_l would generate Landscape format PostScript files.
Initial suggested value: 0.0.
.ARC line list file. This will normally be Angstroms but may be
changed by supplying an appropriate .ARC file.
IDX_ parameters provide a method of indexing reduction database arrays by order number and frame number.
A non-zero value for IDX_NREF_FRAME causes processing of the selected frame only.
IDX_NREF_FRAME=0 causes automatic looping through all values for the ‘index’, i.e., processing of all arc frames.
The IDX_ parameters provide a method of indexing reduction database arrays by order number and frame number.
A non-zero value for IDX_NREF_ORDERS causes processing of the selected order only.
IDX_NREF_ORDERS=0 causes automatic looping through all values for the ‘index’, i.e., processing of all orders.
Many options are available in this mode to allow the tuning of the fitted wavelength polynomial, addition and deletion of features etc.
If ID_INTERACTIVE is not set then the echomop will try to identify arc-line features automatically. For orders with enough features (more than 12 lines) this will usually be successful.
A useful strategy is often to do an automatic run first, and then manually examine the fitted features to verify their suitability.
< function in echomop).
This will usually be the Object frame, however the Standard-Star frame may also be extracted. In cases where there are multiple frames of the same Object, then they may each be extracted by re-running this task and specifying a different frame each time (assuming the multiple data frames are perfectly registered).
.ARC line list file. This will normally be Angstroms but may be
changed by supplying an appropriate .ARC file.
.ARC line list.
.ARC line list.
These will usually be arc spectra. The default is a single frame, but it is possible to use two frames which bracket the object exposure in time. In this case it is assumed that the shift between exposures has been due to smooth variation, an assumption which it is usually not possible to verify.
Note: the names of all frames must be supplied in response to the ARC prompt when locating the
reference features. For example:
ARC=<frame1>,<frame2>
The available functions are POLY (the default) and SPLINE.
Note that the required maximum number of coefficients is different in each case. For POLY the
maximum number of coefficients corresponds to the number of degrees plus one, so for a cubic
use coeffs=4. For a SPLINE the maximum number of coefficients corresponds to 2x(knots+7), so
for 3 knots use coeffs=20. If you are using SPLINE fits then you should ensure that the
value of TUNE_MAXPOLY is set accordingly, as well as using an appropriate value for
TRC_NPOLY.
Initial suggested value: POLY.
The options are:
D — Calculate/edit Dekker limits.
O — Calculate/edit Object profile.
S — Calculate Flux standard profile.
A — All, do each of the above in turn.
Normally this parameter would be set to A, and the program run automatically. The profile and calculated limits examined as the program ran, and then re-run interactively with one of the above options.
It is only needed in cases of severe geometric distortion and is organised such that pixel columns correspond to linear wavelength bins in a given order.
Options are:
The following types are recognised:
— EXTOBJ, the unscrunched extracted object orders.
— EXTARC, the unscrunched extracted arc orders.
— SCROBJ, the scrunched extracted object orders.
— SCRARC, the scrunched extracted arc orders.
— OSPECT, the merged object spectrum.
The default is EXTOBJ.
— OBJ, the object.
— STAR, the reference star.
— ARC, the arc lamp.
YES if the current frames wavelength range is to be used (all orders) to generate a
full range wavelength scale to scrunch into; i.e., all orders are scrunched into the same bin size,
but the no of X-axis bins in each order will vary. Set to NO if each order is to be scrunched to its
own scale. This retains the X-dimension of the orders, but each order will be scrunched into a
slightly different bin size.
The available functions are MEAN (the default), POLY and SPLINE.
Note that the required maximum number of coefficients for POLY and SPLINE are different. For
POLY, the maximum number of coefficients corresponds to the number of degrees plus one, so
for a cubic use coeffs=4. For a SPLINE the maximum number of coefficients corresponds to
(knots+7)*2, so for 3 knots use coeffs=20. If you are using SPLINE fits then you should ensure that
the value of TUNE_MAXPOLY is set accordingly, as well as using an appropriate value for
TUNE_SKYPOLY and/or TUNE_SKYXPLY.
Initial suggested value: MEAN.
This should be a frame which clearly delineates the inter-order gap and the full extent of the slit. In general the arc-lamp frame is used, but a flat-field may also work.
SOFT is specified, the plot is written to the device defined as the current softcopy
device.
The ’_STK.sdf’ part of the filename should not be given.
The default value is ECHOMOP.
Initial suggested value: 0.0.
The following options are available:
ICUR) coordinates are supplied.
Plus one of the following mandatory selections:
For example:
The default mode is C which provides a reasonable balance of probability-of-a-good-trace vs.
time-taken. The order of specification is significant and must obey the following
rules:
— Optional U or R followed by
— Optional T and/or A followed by
— One of B, E, C or G.
The available functions are POLY (the default) and SPLINE.
Note that the required maximum number of coefficients is different in each case. For POLY the
maximum number of coefficients corresponds to the number of degrees plus one, so for a cubic
use coeffs=4. For a SPLINE the maximum number of coefficients corresponds to 2x(knots+7), so
for 3 knots use coeffs=20. If you are using SPLINE fits then you should ensure that the
value of TUNE_MAXPOLY is set accordingly, as well as using an appropriate value for
TRC_NPOLY.
Initial suggested value: POLY.
Set for interactive order fitting/clipping.
If not set, order fitting/clipping proceeds automatically. When automatic fitting is engaged there are two parameters which may be adjusted to optimise the performance of the program. These are: TUNE_CLPMXDEV, which defines the clipping point, and TUNE_CLPBY, which is the number of points to clip in each iteration.
The default value is NO.
The type of function to be used is set using the parameter TRCFIT. (Default value is POLY for
polynomial.)
When TRCFIT=POLY then the number of coefficients should be set to degree of polynomial +
1.
When TRCFIT=SPLINE the number of coefficients should be set to 2x(number of knots +
7).
For example, setting TRC_NPOLY to 4 would permit the fitting of polynomials of the form:
constant + ax + bx squared + cx cubed Initial suggested value: 4.
YES if the order consistency is to be checked interactively. Automatic checking is
performed otherwise.
When automatic checking is engaged there are two parameters which may be adjusted to optimise the performance of the program.
These are TUNE_CLPMXDEV which defines the clipping point, and TUNE_CLPBY which is the number of points to clip in each iteration.
When TRC_VETO is set, the option to reject a re-fitted order at each iteration is available.
It is strongly recommended that this option is not used until you have a clear understanding of the algorithm used by the consistency checker, as it is possible to ‘drive’ the fit off by injudicious rejection of the re-fits proposed by the program.
Initial suggested value: 1.
If set to YES, then the vacuum wavelengths are converted using the procedure:
Only the wavelength scales in the output file are effected; the reduction database wavelength scales remain unchanged.
Initial suggested value: NO.
YES if automatic archiving of results files is to be performed. The archive is
maintained by a central site and will support query facilities. A 2-year data protection
scheme is applied, so your spectra may not be accessed by other parties until 2 years
after its inclusion in the archive. See HELP ECHOMOP ARCHIVING for a more detailed
discussion.
The default value is NO.
When not set, the program allows interactive selection of the order centres on a graphic display. This is useful if you wish to select a subset of the available orders in the frame, or to review the order centres found by the program.
The default value is NO.
Any options required must be supplied on the command line or the you will be prompted as usual. For example:
TUNE_AUTOMATE=1,2,3,EXIT
would run the first 3 processing options and then exit.
If EXIT is not given as an option, echomop will start the menu task ECHMENU when the list of options specified by TUNE_AUTOMATE has been exhausted.
YES for batch-mode operation.
The default value is YES.
Initial suggested value: YES.
The reduction database should have been created by echomop previously. Objects which would normally be calculated by echomop will instead be copied from the named file. This allows the cloning of sections of a reduction amongst multi-frame datasets.
The TUNE_CLPBY most-deviant points are clipped before a re-fit is attempted, this process is then repeated until either: all points have deviations of less than TUNE_CLPMXDEV from the polynomial; or, so many points have been clipped that a meaningful fit can no longer be attempted. In the latter case the order will be automatically disabled. You may manually re-fit the trace if the order is still to be processed.
The default value is 1.
It is expressed in units of pixels and will usually be a fraction of a pixel.
Any traced point which is further than this limit from the fitted polynomial (for that X-coordinate) is liable to be clipped. Automatic clipping only stops when either all points have deviations of less than TUNE_CLPMXDEV from the polynomial; or when so many points have been clipped that a meaningful fit can no longer be attempted. In this latter case, the order will be automatically disabled. You may re-fit manually if the order is still to be processed.
The default value is 0.5.
The default value is 0.5 pixels.
YES then the optimal extraction will attempt to identify cosmic rays
by examining the local variance. This option will only function when full optimal
extraction has been selected using EXTRACT_MODE=O. It will not normally be used as more
flexible tasks for cosmic-ray removal are provided as separate programs. See HELP
ECH_DECOSMIC.
When set, cosmic-ray pixels are replaced in the output image by interpolated values.
When not set, cosmic-ray pixels are marked with a flag value.
The default is NO.
You may use an estimate, or else a value can be generated by an automatic modeller.
If no estimate is available, then zero should be used. The program will then display a histogram of the ratio of original divided-by median image and allow manual selection of a clipping point. The number of pixels to be clipped (flagged as cosmic rays) will then be shown.
This will not usually be necessary and is provided for the case where only a faint, contaminated object frame is available for order tracing.
Initial suggested value: NO.
A non-zero value for this parameter will over-ride the automatic dekker extent calculated by ech_spatial.
TUNE_DEKABV can be used when the dekker extent and position relative to the order traces are well known. Typically, it will be used when the first of a set of similar frames has been manually extracted and automation of the extraction of the remaining frames is required.
Initial suggested value: 0.
A non-zero value for this parameter will over-ride the automatic dekker extent calculated by ech_spatial.
TUNE_DEKBLW can be used when the dekker extent and position relative to the order traces are well known. Typically, it will be used when the first of a set of similar frames has been manually extracted and automation of the extraction of the remaining frames is required.
Initial suggested value: 0.
This is the percentage (of maximum intensity) below which the spatial profile is to fall before being classified as outside the dekker. In general an Arc frame will be used to establish the dekker limits, and a value of 80% will suffice.
The default value is 0.8 (i.e. 80%).
In the event of a problem mail this file, with details of the problem, to the Starlink Software
Librarian at: starlink@jiscmail.ac.uk.
The default value is 0.9 (i.e. 90%).
The trace frame is checked and other frames are then assumed to have identical ‘bad’ features.
Set if the median value is to be used.
This applies only when TUNE_FFNXPLY=1 (no polynomial model), in which case the local median or mean is taken over the nearest TUNE_FFLSMP pixels. This is useful for cases where there is significant small-scale structure in the flat field.
Initial suggested value: NO.
This parameter is only used when the X-polynomial degree, TUNE_FFNXPLY=1, which implies a constant flat field.
The parameter TUNE_FFLMED controls whether a local mean or median value is used.
Initial suggested value: 10.
A model of the flat field is made by fitting a polynomial at each X- and Y-increment. These polynomials are then used to calculate the balance factors at the ’Object’ pixel positions.
Note: If the parameter is set to zero then no modelling in the X-direction will be performed. If the parameter is set to 1, then no polynomials are used (they would be constant), but the balance factors are calculated using the local mean value based on a TUNE_FFLSMP pixel sample. This option should be used when under-sampling etc. is preventing the fitting of polynomials at all.
The default value is 1.
The model of the flat field is made by fitting a polynomial at each X- and Y-increment. These polynomials are then used to calculate the balance factors at the ’Object’ pixel positions.
The model of the flat field is made by fitting a polynomial at each X- and Y-increment. These polynomials are then used to calculate the balance factors at the ’Object’ pixel positions.
The default value is 0.
The model of the flat field is made by fitting a polynomial at each X- and Y-increment. These polynomials are then used to calculate the balance factors at the ’Object’ pixel positions.
Set if subsampling is to be done.
This parameter is only used when polynomial fitting in the X-direction is enabled (by setting TUNE_FFNXPLY>1) and causes the program to subsample (10 samples per-pixel) the spatial profile during fitting. This is useful when the balance factors ‘outside’ the dekker, or at the very edge of the dekker (where the intensity changes quickly from pixel-to-pixel) are needed.
Initial suggested value: NO.
Any pixels which deviate by more than TUNE_FFTHRESH sigma from the fitted polynomial model will be clipped and the model re-fitted.
The polynomial model of the flat field is made by fitting a polynomial at each X- and Y-increment. These polynomials are then used to calculate the balance factors at ’Object’ pixel positions.
The default value is 10.0 sigma.
YES if a multi-fibre object frame.
YES if the program is to conserve flux. If not set, it will maintain the mean height of the
data.
Initial suggested value: 8.
Initial suggested value: 30.
The default value is 6.
giving a fractional measure of the difference. For example: difference=0.01 indicates that the two quantities differ by approximately 1% of their average magnitude. Values much larger than 0.1 are likely to generate a lot of coincidence matches; values less than 0.01 may well miss ‘good’ matches in less-than-ideal data.
The default value is 0.03.
Higher values are tried automatically by the program if no solution can be found. The number of neighbours considered is increased monotonically until it reaches the maximum of TUNE_IDMDLT, when the program gives up.
The default value is 3.
The default value is 10.0, thus all features which are at least one tenth as strong of the strongest
feature will be eligible for identification.
For POLY trace fits, the order of the polynomials used for each fit may be varied up to this limit,
but may not be set to higher values.
For SPLINE trace fits this is the maximum number of fit parameters which can be used. A high
vale may be needed. The number of parameters required is 2x(knots+7).
The default value is 50, which should be acceptable in most cases.
YES if the scrunched object spectrum is to be merged into the 1-D spectrum. The default is
to zero the 1-D spectrum before output. Setting TUNE_MERGE=YES allows the spectra from
multiple frames to be co-added into the same output spectrum using a weighted
merge.
The algorithm is the same as used in the fiGARO ECHMERGE program.
Initial suggested value: NO
Only those pixels with intensities greater than TUNE_MINCR will be flagged. This provides a limit for the case where a cosmic-ray identifier becomes over enthusiastic due to the nature of the frame.
Normally this parameter should be set to zero, and therefore, have no effect.
Initial suggested value: 0.0.
The default value is 10.
This parameter is used to dimension objects in the reduction database and should only need to be altered when dealing with objects where the spatial extent is large.
Note that increasing this parameter leads to a substantial increase in the size of a reduction database, so it should be set to as small a value as sensible for your data.
If the X-dimension is less then 500 every X-increment will be sampled.
The default value is 500.
YES if there is no arc frame to use for wavelength calibration. This may also be specified
by replying NONE when prompted for the name of the arc frame.
Initial suggested value: NO.
This may also be setting the parameter FFIELD to NONE when prompted for the name of the
flat-field frame.
In the absence of a flat-field frame, all balance factors are set to unity.
Initial suggested value: NO.
A non-zero value for this parameter will over-ride the automatic object extent calculated by ech_spatial.
TUNE_OBJABV can be used when the object extent and position relative to the order traces are well known. Typically, it will be used when the first of a set of similar frames has been manually extracted and automation of the extraction of the remaining frames is required.
Initial suggested value: 0.
A non-zero value for this parameter will over-ride the automatic object extent calculated by ech_spatial.
TUNE_OBJBLW can be used when the object extent and position relative to the order traces is well known. Typically, it will be used when the first of a set of similar frames has been manually extracted and automation of the extraction of the remaining frames is required.
Initial suggested value: 0.
The default value is 0.
The polynomial model of the object’s spatial profile is made by fitting a polynomial at each X-sample. These polynomials are then used to estimate the profile at each spatial sample of the object.
Any pixels which deviate by more than TUNE_OBJRTHR sigma from the fitted polynomial model, will be clipped, and the model re-fitted. The polynomial model of the object profile is made by fitting a polynomial at each X-increment. These polynomials are then used to estimate the signal at each spatial sample of the object.
Initial suggested value: 5.0.
When set to a non-zero value, after each TUNE_PAGE lines of non-menu output the program will pause and wait for a key to be hit before continuing. A variety of options are available at this point:
| D,d,uparrow,KP8,p,P,b,B | Scroll back one line. |
| U,u,downarrow,KP2,f,F,n,N | Scroll forward a line. |
| space | Resume output. |
| return | Output next line only. |
| pageup | Previous full screen. |
| pagedown | Next full screen. |
The program maintains a buffer of lines output to the screen; you may move around using the keys mentioned above. Only the space or return keys will resume regular program output. If the parameter value is not positive then no paging is done.
Initial suggested value: 0.
The default value is NO.
This parameter should only need to be changed when processing frames with a very large number of pixels sampling the object in the spatial direction.
The default value is 301 subsamples, which will normally be more than is needed.
The default value is 101.
If set, the modelling will not take place and the balance factors will simply be copied from the supplied frame. This should be used for cases where the polynomial model cannot generate appropriate values.
Note: The removal of the blaze function is a separate operation and will require the use of the original raw flat-field frame (or equivalent).
The default value is YES.
This enables fast reduction by automatically tailoring parameters. In particular: all fits are clipped automatically; no flat fielding is performed; and a simple summed extraction is used.
Prior to quick-look extractions the object extent should be specified by the parameters TUNE_OBJBLW and TUNE_OBJABV.
Initial suggested value: NO.
It will usually be left blank to indicate default reporting to the terminal/batch log file. Options include:
| F | Specifies full reporting of task/module entry. |
| L | Specifies duplicate log of reports to disk file |
| (ech_report.log in the current directory). | |
| B | Brief reporting, task starts only. |
| S | Silent, no terminal reports at all. |
| E | Error reports to override other setting (BS). |
| P | Print log file on completion (only with L). |
Specifiers may be concatenated in any order, thus
BELP
would specify minimal reporting, except for task startup and error reports. Reports to be copied to a log file and printed upon completion.
Initial suggested value: ’ ’.
The default value is NO.
The default value is 1.25.
Flagged pixels do not contribute to the extraction in any way. Only pixels with intensities less than the TUNE_SATRTN level will be used by subsequent tasks.
Initial suggested value: 1.0E20.
No facility is provided for calculating the appropriate fraction but it will usually be estimated as the the ratio of elapsed times between the object frame, and the before/after arc frames (assumes a constant ‘drift’ of the instrumental setup).
If not equal to zero, selects mode 1, in which the input bins are transferred to the output bins in accordance with the wavelength scale(s).
YES then weights are modified to include uncertainty in sky fit
optimal variances equivalent to maximising:
where:
is the weight on pixel ,
is the variance on the value of pixel ,
is the covariance of the skyfit between pixels and ,
is the normalised profile at pixel .
This reduces to solving:
where:
for , and
.
The algorithm used is:
— Median filter profile and multiply by 1.05
— Calculate the threshold point: Min + (Max-Min) x TUNE_SKYHILIM
— Take the minimum of the two quantities
This minimum is used as the maximum sky intensity allowed along the spatial profile.
The default value is 0.5 (i.e. 50%).
This is only used when a wavelength-dependent model is being used to predict sky intensities using whole-order polynomials. In this case, any pixel whose intensity differs from the fit by more than sigma TUNE_SKYLTHR will be treated as a sky-line candidate, and its actual intensity entered into the model in place of the fit prediction. The same procedure is also applied to the pixels’ TUNE_SKYLINW nearest neighbours to ensure that the entire sky line is entered into the model even when its wings are within the sigma constraint described above. This parameter only has effect when TUNE_SKYXPLY>0.
The default value is 5 pixels.
In this case, any pixel whose intensity differs from the fit by more than TUNE_SKYLTHR x fit-sigma will be treated as a sky-line candidate, and its actual intensity entered into the model in place of the prediction. The same procedure is also applied to the pixels TUNE_SKYLINW nearest neighbours to ensure that the entire sky line is entered into the model even when its wings are within the sigma constraint described above.
Initial suggested value: 3.0.
The default value is TUNE_SKYPOLY=0 to obtain sky averaging (i.e., sky model is constant at each order/column).
Set TUNE_SKYPOLY=-1 when there is no sky to be modelled (i.e., sky model is zero everywhere).
Initial suggested value: 0.
Initial suggested value: 5.0.
YES if simulations are to be used to estimate the variance of the sky using polynomial fits
in the wavelength direction.
These simulations are CPU intensive.
The default value is NO.
This facility will be most useful when the sky is very uniform, or possibly when trying to model the scattered light in the inter-order region. The intensities predicted by the wavelength-dependent model are used as input to the spatial model only when they fall within a user-defined distance from the observed value, and when they are not possible sky-line features.
(See also TUNE_SKYLINW and TUNE_SKYLTHR).
The default value is 0.
The default value is 0.05.
The threshold is the degree to which off-trace pixels must fall below the pixel at trace centre (in the spatial direction) in order to be considered as ’off-order’. I.e., the trace width is estimated by stepping pixel-by-pixel above and below the trace until the intensity falls below (1.0 - TUNE_TWTHR) x intensity-at-trace-centre.
The default value is 0.95 (95%).
YES if the program is to make use of the ‘Astronomy & Astrophysics Abstracts’ system of
information categorisation. This consists of a set of numbers corresponding to topics in
Astronomy. If TUNE_USEAAA is active then you will be shown the currently recognised categories
and asked to select the one most relevant to the data being reduced. This parameter is only used
in conjunction with the archiving facility, when TUNE_ARCHIVE=YES. The prompting
process may be overridden by supplying a valid AAA code using the parameter
TUNE_AAACODE.
Initial suggested value: YES.
This is set low (0.2=20%) by default for speed.
Very poor data may benefit from an increase in this parameter, which provides a more accurate sampling of the object profile, and thus more accurate extraction weights. The central fraction from each order is sampled and the average profile over all orders is calculated.
A special mode is selected by setting TUNE_USE_NXF>=1, which makes the sampling independent for each order. In this case the fraction of the X-dimension of the frame used is TUNE_USE_NXF-1, for example, TUNE_USE_NXF=1.2 would select individual-order profiling using 20% of each order.
The default value is 0.2 (i.e. 20%).
This should be necessary when only one frame is available for tracing the orders, and this frame is of poor quality.
Some care must be taken to ensure that the sampling box is kept sufficiently small that all samples come from the same order for each sampling-box position.
YES if automatic Y-blaze correction is to be used. This correction works by examining the
scrunched orders and re-scaling them so that the extremes of adjacent orders match up. This
should only be used when no flux calibration is available and there is a real need
for it. It requires that the blaze fit be scrunched before being applied and therefore
necessitates a blaze-fit followed by the scrunch option. The blaze correction is only
applied to scrunched (and merged) orders in this case. When disabled then blaze
fitting is applied to the extracted orders directly and these may be scrunched later if
required.
Initial suggested value: NO.
YES if median is to be used.
YES if you want to interactively control the 2-D polynomial fitting of the distortion
corrections. Polynomials are fitted in two perpendicular directions (rows and columns)
and generate the point-to-point predicted deviation from wavelength fit for each
pixel centre. Interactive operation allows you to view an exaggerated version of the
fitted points, and to selectively clip points on the basis of their deviations from the
fit.