aboutsummaryrefslogtreecommitdiff
path: root/noao/digiphot/daophot/doc/specs/daophot.spc
diff options
context:
space:
mode:
Diffstat (limited to 'noao/digiphot/daophot/doc/specs/daophot.spc')
-rw-r--r--noao/digiphot/daophot/doc/specs/daophot.spc1047
1 files changed, 1047 insertions, 0 deletions
diff --git a/noao/digiphot/daophot/doc/specs/daophot.spc b/noao/digiphot/daophot/doc/specs/daophot.spc
new file mode 100644
index 00000000..221630d5
--- /dev/null
+++ b/noao/digiphot/daophot/doc/specs/daophot.spc
@@ -0,0 +1,1047 @@
+.help daophot Sep87 "Crowded Field Stellar Photometry'
+.sh
+1. Introduction
+
+ The DAOPHOT package will provide a set of routines for performing
+stellar photometry on crowded fields in either interactive or batch mode.
+DAOPHOT works by fitting an empirical point spread function (PSF)
+to each object in the field
+allowing for overlap of closely spaced images. This document presents the
+the requirements and specifications for the package and describes some of
+the algorithms to be used. Most of the algorithms are described in
+the original article by Peter Stetson (1987 P.A.S.P. 99,191).
+
+.sh
+2. Requirements
+.ls 4
+.ls (1)
+The tasks in the DAOPHOT package shall take as input an IRAF imagefile
+containing two-dimensional image data which has been corrected for
+pixel to pixel gain variations, high frequency variations in the background,
+any nonlinearitys in the data except for those which can be specified as
+a lower and/or upper bound,
+and any other instrumental defects affecting the intensity value of an
+individual pixel. However, it shall be possible to exclude bad pixels,
+rows or columns from analysis by DAOPHOT routines in a very crude manner.
+.le
+.ls (2)
+The tasks in the package which produce tabular output shall use the
+SDAS Tables for their output and those tasks which read output from other
+DAOPHOT tasks will be able to read SDAS Tables. In the future the input/output
+shall make use of the DBIO package.
+.le
+.ls (3)
+The DAOPHOT package shall work in conjunction with the APPHOT package produced
+at NOAO. DAOPHOT will not have any provision to do aperture photometry of its
+own. The output format from DAOPHOT tasks will be consistent with APPHOT.
+.le
+.ls (4)
+Given as input a reduced two-dimensional image which has been processed by the
+APPHOT package, the DAOPHOT package shall be able to perform the following
+functions:
+.ls 4
+.ls o
+Interactively define a PSF for the data frame. The PSF will be defined
+empirically from one or more stars in the field. The task to determine the
+PSF shall be interactive and the user shall be able to use a
+graphics terminal and/or an image display device to select the stars which
+will make up the PSF. The user will be able to evaluate the PSF through
+different means including contour plots, 3-d mesh plots, and displaying the
+PSF on an image display device.
+The user shall be able to "mask" out parts of the PSF which may be contaminated
+by nearby stars, bad pixels etc. Only the non-masked portions of the PSF will
+be used in the fitting routines.
+.le
+.ls o
+Fit the PSF simultaneously to groups of stars in the image frame whose
+images overlap to some degree. The parameters in the fit shall include the
+the object brightness, X and Y position of the star and potentially the sky
+background. The sky shall be able to be specified as either a flat uniform
+background or a simple tilted planar sky. The photometry routines shall
+produce realistic errors in the photometry assuming that realistic numbers
+for the characteristics of the data are input.
+.le
+.ls o
+Subtract the fitted stars from the data frame to produce a subtracted
+image for further analysis.
+.le
+.ls o
+Add artificial stars to the data frame in order to check accuracy and
+completeness in the photometry. The user shall have control over the
+number of stars added, the brightness range, the area of the image to contain
+the added stars and the noise characteristics of the added stars.
+.le
+.le
+.ls (5)
+The DAOPHOT package shall include tasks to inspect and edit the results from the
+photometry routines. These shall include tasks such as interactively
+rejecting particular stars from the results,
+producing plots of errors versus brightness, errors versus position etc.
+.le
+.ls (6)
+The DAOPHOT package shall provide utility packages to handle the output
+data from the fitting routines. These shall include such tasks as
+aperture growth curves, photometric calibrations, color-magnitude and
+color-color diagrams.
+.le
+.ls (7)
+The DAOPHOT routines shall optionally keep a history file to keep
+track of the processing done on the images. This will include the values of
+various parameters used in the various tasks of the DAOPHOT package.
+.le
+.ls (8)
+The tasks shall be able to be run in batch mode as well as interative
+mode. In batch mode use of a graphics terminal or image display shall not
+be required.
+.le
+.ls (9)
+The DAOPHOT package shall be written in the SPP language in conformance with
+the standards and conventions of IRAF. The code shall be portable and
+device independent.
+.le
+.le
+.sh
+2.1 Limitations of the Initial DAOPHOT Package
+
+The DAOPHOT package shall perform PSF fitting photometry with the following
+restrictions:
+.ls
+.ls (1)
+The PSF used will be determined empirically and analytic specification of
+the PSF will not be possible. This restricts the use of DAOPHOT to image
+data which is not too badly undersampled.
+.le
+.ls (2)
+There will be an upper limit to the number of stars for which the PSF will
+be fit simultaneously. The initial version of DAOPHOT will have this limit
+set to 60 stars.
+.le
+.ls (3)
+The initial version of DAOPHOT will not have the sky included as a parameter
+in the fitting routines.
+.le
+.ls (4)
+Initially the use will not be able to mask out bad portions of the PSF for
+fitting.
+.le
+.le
+
+.sh
+3. Specifications
+
+ The DAOPHOT package performs stellar photometry on digital data, maintained
+as IRAF image files. DAOPHOT performs this photometry by fitting the PSF
+to the stellar images in the image file. DAOPHOT works by fitting the PSF to
+a maximum number of stars simultaneously thus allowing for overlapping images.
+Input to the package consists of an imagefile and the output from the APPHOT
+package, which contains simple aperture photometry for the objects which have
+been identified in the image frame, and numerous parameters controlling the
+analysis algorithms. Output from the analysis tasks consists of tabular data
+containing the results of the analysis routines. The output will be in the
+form of SDAS tables and will thus be able to be manipulated by various
+other utility tasks available in IRAF.
+
+The CL callable part of the DAOPHOT package consists of the following routines:
+
+.ks
+.nf
+ addstar -- adds synthetic stars to an image file
+ allstar -- fits multiple, overlapping PSFs to star images
+ *calibrate -- apply photometric calibration
+ *cmd -- color-magnitude, color-color diagrams
+ daopars -- DAOPHOT pset parameters
+ examine -- interactively examine/edit photometry results
+ group -- divides stars into natural groupings
+ *growth -- aperture growth curves <--> PSF magnitudes
+ peak -- fit PSF to single stars in an image file
+ psf -- interactively construct a PSF for the frame
+ nstar -- fits multiple, overlapping PSFs to star images
+ seepsf -- converts a PSF file into a IRAF image file
+ select -- selects natural groups with a certain range of sizes
+ substar -- subtract fitted profiles from an image file
+.fi
+.ke
+
+There are routines available in other IRAF/STSDAS tasks for manipulating
+SDAS Tables or DBIO. The capabilities inside the DAOPHOT are specifically
+suited to dealing with large tables of results from these photometry routines.
+
+.sh
+3.1 Standard Analysis Procedures
+
+ Before performing DAOPHOT photometry one must perform certain other tasks
+beforehand. This includes using the APPHOT package to produce an object list
+and aperture photometry for objects in this list. The DAOPHOT package contains
+an additional object finder but one must use APPHOT to obtain the aperture
+photometry results. The standard analysis procedure, including APPHOT, is as
+follows:
+.ls
+.ls (1)
+Use an object finder to produce a list of object coordinates. This may be done
+in many ways:
+.ls
+.ls o
+By using the interactive cusrsor routines available elsewhere in IRAF and
+redirecting the output into a list file.
+.le
+.ls o
+By transforming an existing list using an existing IRAF task or the OFFSET
+task in the DAOPHOT package.
+.le
+.ls o
+By using an automatic object finding procedure such as the one available
+in the APPHOT package or the one in the DAOPHOT package.
+.le
+.ls o
+By any other program which generates a list of objects in suitable format (SDAS
+Tables) for input to the APPHOT routines.
+.le
+.le
+.ls (2)
+The APPHOT package is run to measure the objects identified in the above
+step. One should refer to the APPHOT documentation to understand the
+algorithms and procedures which are used in APPHOT.
+.le
+.ls (3)
+One needs to set up the parameters in the analysis routines for this particular
+image file. OPTIONS allows you to set such parameters as the number of
+electrons/ADC, the fitting radius, and the radius within which the PSF
+is defined.
+.le
+.ls (4)
+The next step is to produce a PSF for the image file currently being processed.
+In crowed fields this is a tricky, iterative procedure which should be done very
+carefully. This is best done using a graphics terminal and/or an image display
+device.
+.le
+.ls (5)
+If one plans on using NSTAR, then the GROUP task must be run. This task
+divides the stars in the output from the APPHOT into natural groups. The size
+of the groups produced depends upon how crowded the field is and what degree of
+overlap of the images one considers.
+.le
+.ls (6)
+Use either NSTAR, if you have grouped the objects using GROUP, or
+ALLSTAR which will dynamically group the stars as the image file is
+processed. These routines will produce the objects' positions and
+intrumental magnitudes by means of multiple-profile fits.
+.le
+.ls (7)
+Use SUBSTAR to subtract the fitted profiles from the image file, thus producing
+a new image file containing the fitting residuals. This will usually contain
+many stars which were missed in the original identification because they lie
+in the wings of brighter objects.
+.le
+.ls (8)
+One now basically runs through steps (1) - (6) one or more times,
+merging the identified object lists each time to produce a master object list,
+until one is satisfied with the final results. There are many subtlties in this
+procedure which are described in the DAOPHOT User's Manual.
+.le
+.ls (9)
+After obtaining the photometry results one may edit the results by throwing out
+those results which do not meet certain criteria. EXAMINE is an interactive
+task which allows the user to examine the results for each individual object
+in the list and either accept or reject that object. There are also routines
+available for courser rejection of results, e.g. reject all objects with
+errors larger than 0.2 magnitudes.
+.le
+.ls (10)
+One may wish to use the tasks to plot up color-color or color-magnitude
+diagrams. Other general purpose list processing tools available in
+IRAF/SDAS may also be used for analysis of DAOPHOT output.
+.le
+.le
+
+.sh
+3.2 The ADDSTAR Task
+
+ The function of ADDSTAR is to add synthetic stars to the image file.
+These stars may be placed randomly by the computer, placed with a certain
+distribution as specifed by the user or at predetermined locations specified
+by the user. Likewise the brightness of these added objects may be completely
+random or may follow a specified distribution.
+
+Objects are added by taking the specified PSF, scaling it, and moving it
+to the desired location. ADDSTAR will also add Poisson noise to the star
+images to make them more realistic.
+
+.sh
+3.2.1 ADDSTAR Parameters
+
+ ADDSTAR has several parameters which control the addition of stars
+into a image file. All data dependent parameters are query mode to ensure
+that they get set properly for the particular image under consideration.
+The data independent parameters are hidden mode, and are given reasonable
+default values. The names, datatypes, and default values of the ADDSTAR
+parameters are shown below.
+
+.ks
+.nf
+Positional or query mode parameters:
+
+ input_image filename
+ output_image filename
+ minmag real
+ maxmag real
+.fi
+.ke
+
+.ks
+.nf
+List structured parameter (filename may be given on command line):
+
+ add_data *imcur
+.fi
+.ke
+
+.ks
+.nf
+Hidden Parameters:
+
+ daopars pset "daophot$daopars.par"
+ nstar integer 100
+ nframe integer 1
+ xmin integer 1
+ ymin integer 1
+ xmax integer NX
+ ymax integer NY
+ verbose boolean false
+.fi
+.ke
+
+The function and format of each of these parameters is explained in
+more detail below.
+
+.ls
+.ls 16 input_image
+The name of the image or image section to which artificial stars will be added
+.le
+.ls output_image
+The name of outout image which will contain the added stars.
+.le
+.ls minmag
+The minumum magnitude of artificial star to add to the data. The magnitude
+scale is set by the magnitude of the PSF.
+.le
+.ls maxmag
+The maximum magnitude of artificial star to add to the data. The magnitude
+scale is set by the magnitude of the PSF.
+.le
+.ls add_data
+This parameter is used to specify a file as input to the ADDSTAR task. This
+file should contain centroid positions and magnitudes for the stars you
+want to add. It is possible to specify the positions of the added stars
+interactively with the image display by setting this parameter to *imcur.
+In this case the user is prompted for the magnitude of each star to be added.
+If this parameter is the null string then the stars are added in a random
+fashion by the ADDSTAR routine.
+.le
+.ls nstar
+The number of artificial stars to add to the input image file.
+.le
+.ls daopars
+This is the name of a file containing parameters which are common to
+many DAOPHOT tasks. This pset parameter serves as a pointer to the external
+parameter set for the DAOPHOT algorithms. The parameters contained in this
+pset and their function are described in section 3.6.1.
+.le
+.ls nframe
+The number of new image files to create. If this parameter is greater
+than one then the new image files will use the output image name as
+a root and produce image files with '.xxx' appended to the root, where
+xxx will range from 001 to nframe. If nframe is one then the output image
+name will be used as is.
+.le
+.ls xmin, ymin, xmax, ymax
+These define the subsection of the image in which to add the artificial
+stars. The default is to add artificial stars to the complete image.
+.le
+.ls verbose
+Controls the amount of output from the ALLSTAR function. The default is
+to have minimal output to STDOUT.
+.le
+.le
+
+.sh
+3.2.2 ADDSTAR Output
+
+ The output of ADDSTAR consists of two parts, an image file and an
+output SDAS Table. The image file is a copy of the input image file but
+with the artificial stars generated by ADDSTAR added. The output table
+contains the x,y position and magnitude of each of the added stars. When the
+nframe parameter is set greater than one then there will be nframe pairs of
+output files generated.
+
+.sh
+3.3 ALLSTAR
+
+ ALLSTAR fits multiple, overlapping point-spread functions to stars images
+in the input image file. It uses as input the results from APPHOT and an
+input PSF and will automatically reduce the entire image performing the necessary
+grouping. It will recalculate the grouping after each iteration. ALLSTAR will
+also produce the star-subtracted image file.
+
+.sh
+3.3.1 ALLSTAR Parameters
+
+ALLSTAR has several parameters which control the fitting algorithms. The
+names, datatypes, default values for the ALLSTAR parameters are given below.
+
+.ks
+.nf
+Positional parameters:
+
+ input_image filename
+ photometry filename
+ output filename
+ sub_image filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ daopars pset "daophot$daopars.par"
+ max_group integer 60
+ redeterm_cent boolean true
+ max_crit real 2.5
+ min_crit real 1.2
+ clip_exp integer 6
+ clip_range real 2.5
+ verbose boolean false
+.fi
+.ke
+
+These parameters perform the following functions:
+
+.ls 4
+.ls 16 input_image
+The name of the input image file.
+.le
+.ls photometry
+The name of the input photometry SDAS table. This may contain output from either
+the APPHOT package or from NSTAR or previous ALLSTAR runs.
+.le
+.ls output
+The name of the SDAS table to contain the results of the psf fittting.
+.le
+.ls sub_image
+The name of the output image file which will have all of the fitted stars
+subtracted from it. If this file is the null string then no star-subtracted
+image file will be produced.
+.le
+.ls daopars
+The pset parameter file containing the DAOPHOT parameter set.
+.le
+.ls max_group
+The maximum size group which ALLSTAR will process. The absolute maximum
+is 60 stars.
+.le
+.ls redeterm_cent
+If true then the centers of the stars are redetermined before each
+iteration.
+.le
+.ls max_crit
+The initial value which ALLSTAR uses as the critical separation for
+use in grouping stars together. For groups larger than "max_group" ALLSTAR
+will use progressively smaller values for the critical separation until the
+group breaks up into units containing fewer than "max_group" stars or until
+the value of "min_crit" is reached.
+.le
+.ls min_crit
+The smallest value of the critical separation which ALLSTAR will use in
+grouping stars together.
+.le
+.ls clip_exp, clip_range
+These parameters are used to "resist bad data". These two
+parameters control the weighting of each pixel as a function of it's
+residual from the fit. Clip_range us variable "a" and clip_exp is
+variable "b" in the paper by Stetson (P.A.S.P. 99, 191)
+.le
+.le
+
+.sh
+3.3.2 The ALLSTAR PSF Fitting Algorithm
+
+ The algorithms which ALLSTAR uses to do the psf fitting photometry are
+very nearly the same as those used by NSTAR. One is referred to Stetson,
+P.A.S.P. 99, 191, for the details on the various fitting, star rejection,
+and weighting algorithms used in this task.
+.sh
+3.3.3 The Output from ALLSTAR
+
+ The output from ALLSTAR consists of three parts. There is the output
+photometry results, an SDAS Table, and a subtracted image file. The subtracted
+image file is a copy of the input image file minus the fitted stars.
+
+For each object processed by ALLSTAR there is one row in the output SDAS
+Table. Each measured object will have entries for the following items:
+
+.nf
+ star, x, y, mag, magerr, sky, niter, chi, sharp
+
+where
+
+ star star ID number
+ x,y coordinates of the stellar centroid
+ mag magnitude relative to the magnitude of the PSF star
+ magerr estimated standard error of the star's magnitude
+ sky estimated sky as returned by APPHOT
+ niter number of iterations for convergence
+ chi observed pixel to pixel scatter DIVIDED BY the expected
+ pixel to pixel scatter
+ sharp an index describing the spatial distribution of the residuals
+ around the star. Objects with SHARP significantly greater
+ than zero are extended (possibly galaxies), while objects with
+ SHARP significantly less than zero may be bad pixels or cosmic
+ rays
+.fi
+
+Other noteworthy pieces of information will be stored in the output SDAS
+Table header. This includes such things as the time and date of processing,
+the name of the PSF file, the name of the input photometry file, the
+fitting radius etc.
+
+.sh
+3.4 The CALIBRATE Task
+
+.sh
+3.5 The CMD Task
+
+.sh
+3.6 The DAOPARS Task
+
+ This is a pset-task which is used to describe a particular image file
+for use with the DAOPHOT package. This pset contains parameters which describe the
+data, e.g. the read out noise, the background sky value, the number of photons
+per ADC unit, etc., and also parameters which control the DAOPHOT tasks, e.g.
+the fitting radius to use. The parameters in this pset are used by several
+DAOPHOT tasks, hence their grouping into a pset.
+
+.sh
+3.6.1 daopars Parameters
+
+ The parameters in this task either describe the data in
+a particular image file
+or are parameters which are used by more algorithms in more than one
+DAOPHOT task. The following parameters make up this pset:
+
+.ks
+.nf
+
+ fitrad real 2.5 (pixels)
+ psfrad real 11.0(pixels)
+ phot_adc real 10.0
+ read_noise real 20.0
+ max_good real 32766.
+ min_good real 0.0
+ sky_val real 0.0
+ numb_exp integer 1
+ comb_type string "average"
+ var_psf boolean false
+.fi
+.ke
+
+The function and format of each of these parameters is described below:
+
+.ls 4
+.ls 16 fitrad
+The fitting radius to use in the PEAK, NSTAR, ALLSTAR and PSF tasks. Only
+the pixels within one fitting radius are actually used in the fit. This should
+normally be on the order of the FWHM of the stellar images.
+.le
+.ls psfrad
+The radius of the circle within which the PSF is defined. This should be
+somewhat larger than the actual radius of the brightest star you are
+interested in.
+.le
+.ls maxgood
+The maximum data value in ADC units at which the CCD or other detector
+is believed to operate linearly.
+.le
+.ls mingood
+The minimum data value in ADC units which should be used as "real" data.
+Dead pixels, bad columns etc. in the image file can be excluded from use in
+the analysis by setting this parameters properly. Any data value which
+falls below this minimum is ignored by DAOPHOT tasks.
+.le
+.ls sky_val
+The typical sky brightness in ADC units for the image file. This parameter is
+updated by the SKY task within the DAOPHOT package.
+.le
+.ls phot_adc
+The number of photons per ADC unit of the CCD or other detector.
+.le
+.ls read_noise
+The readout noise in ADC units of the CCD or other detector.
+.le
+.ls numb_exp
+The number of individual exposures which have been combined to produce the
+current image file. This number combined with information on whether the
+exposures were summed or averaged is used to get a better handle on the
+error estimates of the photometry.
+.le
+.ls comb_type
+Describes whether the individual exposures which went into making up this
+image file were "summed" or "averaged"
+.le
+.ls var_psf
+Controls whether the shape of the PSF is to be regarded as constant over the
+complete image file. Slight and smooth variations can be accomodated by the
+DAOPHOT tasks.
+.le
+.le
+
+ These parameters should be initially set by the user before starting any
+analysis with the DAOPHOT package. Each image file may have it's own set of
+parameters and these should be stored in separate pset files.
+.sh
+3.7 The EXAMINE Task
+
+ EXAMINE allows the user to interactively examine the results of the
+DAOPHOT reduction and to accept or reject individual stars. EXAMINE will
+accept as input the output photometry list from either ALLSTAR or NSTAR.
+For each star in the input list the user can examine either a 3-d meshplot
+or a contour diagram of both the input image and the star-subtracted image.
+The results of the photometry for the star under consideration is also
+displayed.
+
+Two output star lists are produced using this task. One is a list
+of stars which have been "accepted" by the user, the other being a list
+of stars which have been "rejected".
+
+If the TV option is selected then both the original image and subtracted
+image are displayed on the "stdimage" and the star under consideration is
+identified. The user has the ability to blink these two frames to
+evaluate the results of the photometry.
+
+This task is controlled via input from the terminal with various keys
+performing a variety of functions.
+
+.sh
+3.7.1 EXAMINE Parameters
+ There are several parameters which control various aspects of the
+EXAMINE task. The parameters control such things as the input photometry
+list, the type of graphical display desired and whether to use the
+display capabilities.
+
+.ks
+.nf
+Query mode parameters:
+
+ phot_list filename
+
+ fwhm real (pixels)
+ threshold real (in units of sigma)
+ output_file filename
+.fi
+.ke
+
+.sh
+3.9 The GROUP Task
+
+ GROUP is used to divide the stars in the image file into natural
+groups prior to analysis with NSTAR. GROUP works on the following
+principle: if two stars are close enough that the light of one will
+influence the profile fit of the other, then they belong in the same
+group.
+
+.sh
+3.9.1 GROUP Parameters
+
+ GROUP only has a few parameters which govern its operation. These
+are:
+
+.ks
+.nf
+Query mode parameters:
+
+ input_image filename
+ psf_file filename
+ crit_overlap real
+ output filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden mode parameters:
+
+ daopars pset "daophot$daopars.par"
+.fi
+.ke
+
+These parameters perform the following functions:
+
+.ls 4
+.ls 16 input_image
+The name of the input image file.
+.le
+.ls psf_file
+The name of the file containing the PSF.
+.le
+.ls crit_overlap
+The "critical overlap" before one star is determined to influence
+another. When GROUP examines two stars to see whether they might influence
+each others' fits, it firts identifies the fainter of the two stars. It then
+calculates the brightness of the brighter star at a distanceof one fitting
+radius plus one pixel from the center of the fainter. If this brightness is
+greater than the "critical overlap" times the random error per pixel, then
+the brighter star is deemed to be capable of affecting the photometry of the
+fainter, and the two stars are grouped together.
+.le
+.ls output
+The name of the SDAS table which will contain the stellar groups.
+.le
+.ls daopars
+The name of of a pset file containing the daophot parameters. The specific
+parameters which are used from this include the following:
+.le
+.le
+.sh
+3.10 The GROWTH Task
+
+.sh
+3.11 The OFFSET task
+
+.sh
+3.12 The PEAK Task
+
+ PEAK fits the PSF to a single star. It is useful for sparsely populated
+image files where the stars of interest are not blended. In this cases aperture
+photometry is often fine and the use of PEAK is of limited interest. This task
+is included in the DAOPHOT package mainly for completeness.
+
+.sh
+3.12.1 PEAK Parameters
+
+ The parameters specific to the PEAK task are used for specifying the
+input and output from this routine. The names of the parameters and their
+functions are:
+
+.ks
+.nf
+Positional or query parameters:
+
+ input_image filename
+ psf_file filename
+ output filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ daopars pset "daophot$daopars.par"
+ verbose boolean false
+.fi
+.ke
+
+.ls 4
+.ls 16 input_image
+The name of the input image file.
+.le
+.ls psf_file
+The name of the input file containing the point-spread function.
+.le
+.ls output
+The name of the SDAS table to contain the output from PEAK.
+.le
+.ls verbose
+If true then PEAK outputs more information about what it is doing.
+.le
+.ls daopars
+The name of a pset file which contains the parameters specific to the
+input image file. The parameters which PEAK uses from this pset include:
+sthe fitting radius, the maximum and minimum good data value and whether
+a variable PSF is to be used.
+.le
+.le
+.sh
+3.13 The PSF Task
+
+ The PSF task is used for obtaining the point-spread function which
+will be used in the rest of the DAOPHOT reductions. DAOPHOT uses an
+empirical point-spread function as opposed to a mathematically defined
+function. The PSF is defined from the actual brightness distribution
+of one or more stars in the frame under consideration. It is stored as
+a two-component model: (1) an analytic Gaussian profile which approximately
+matches the core of the point-spread function, and (2) a look-up table of
+residuals, which are used as additive corrections to the integrated
+analytic Gaussian function.
+
+The brightness in a hypothetical pixel at an arbitrary point within the
+point-spread function is determined in the following manner. First
+the bivariate Gaussian function is integrated over the area of the pixel,
+and then a correction is determined by double cubic interpolation
+within the lookup table, and is added to the integrated intensity.
+
+The PSF is stored as a binary data file and is in a format specific
+to DAOPHOT. The format of this file is very similar to that used by the
+VMS version of DAOPHOT but is stored in binary for compactness.
+A function is provided to take the PSF and convert it
+to a IRAF image file so that it can be manipulated by other IRAF
+tasks.
+
+PSF allows the user to perform most functions from within the interactive
+graphics part of its operation. PSF allows the user to modify the
+perspective of hist mesh plot, the contouring interval, the PSF radius
+etc. from within the PSF interactive graphics.
+
+.sh
+3.13.1 PSF Parameters
+
+ The PSF task has many parameters which specify the input and
+output files as well as specifying other information. These are
+divided in query mode parameters and hidden parameters.
+
+.ks
+.nf
+Positional or query parameters:
+
+ input_image filename
+ phot_list filename
+ psf_stars filename
+ psf_file filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ daopars pset "daophot$daopars.par"
+ verbose boolean false
+.fi
+.ke
+
+.ls 4
+.ls 16 input_image
+The name of the input image file.
+.le
+.ls phot_list
+The name of the input file containing the aperture photometry
+results for this image frame.
+.le
+.ls psf_stars
+The name of file coordinate file containing the list of stars
+to be used as PSF candidates.
+.le
+.ls psf_file
+The name of the output file for storing the PSF.
+.le
+.ls verbose
+If true then PEAK outputs more information about what it is doing.
+.le
+.ls daopars
+The name of a pset file which contains the parameters specific to the
+input image file. The parameters which PEAK uses from this pset include:
+sthe fitting radius, the maximum and minimum good data value and whether
+a variable PSF is to be used.
+.le
+.le
+.sh
+3.14 The NSTAR Task
+
+ NSTAR is one of DAOPHOT's multiple, simultaneous, profile-fitting
+photometry routine. It is similar to ALLSTAR except that NSTAR must have
+the objects grouped (using the GROUP task) and it does not dynamically
+alter the groups while running. NSTAR also does not automatically produce the
+star subtracted image file.
+
+.sh
+3.14.1 NSTAR Parameters
+
+ There are several parameters which control the function of the
+NSTAR task. These are the following:
+
+.ks
+.nf
+Positional or Query Parameters:
+
+ input_image filename
+ psf_file filename
+ group_file filename
+ output_file filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ daopars pset "daophot$daopars.par"
+ verbose boolean false
+.fi
+.ke
+
+.sh
+3.15 The SEEPSF Task
+
+ The SEEPSF task produces an IRAF image file from the given PSF
+file. This allows other IRAF tasks, especially display and plotting tasks,
+to use access the point-spread function. The user has the ability to create
+any size of image from the PSF enlargements being handled by a number of
+different interpolation schemes.
+
+.sh
+3.15.1 SEEPSF Parameters
+
+ The parameters wich control this task are limted. They basically
+control the input, output and size of the image.
+
+.ks
+.nf
+Positional or Query Parameters:
+
+ psf_file filename
+ image_name filename
+ image_size integer
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ interpolation string "nearest"
+ boundary string "constant"
+ constant real 0.0
+ daopars pset "daophot$daopars.par"
+ verbose boolean false
+.fi
+.ke
+
+.ls 4
+.ls 16 psf_file
+This specifies the input PSF file which is to be transformed into an
+IRAF image.
+.le
+.ls image_name
+The name of the output IRAF image.
+.le
+.ls image_size
+The size of the output image in pixels per side. Note that only square PSFs
+and PSF images are alllowed.
+.le
+.ls interpolation
+The type of interpolation to be used in expanding the image. The choices
+are "nearest" neighbor, "linear" bilinear, "poly3" bicubic polynomial,
+"poly5" biquintic polynomial, and "spline3" bicubic spline.
+.le
+.ls boundary
+The type of boundary extension to use for handling references to pixels
+outside the bounds of the input image. The choices are: "constant",
+"nearest" edge, "reflect" about the boundary and "wrap" to the other side
+of the image.
+.le
+.le
+.sh
+3.16 The SELECT Task
+
+ The SELECT task is used to select groups of stars with a particular
+range of sizes from a group file which has been produced by GROUP. This
+task is used when some of the groups in the group file are large than the
+maximum allowed in NSTAR, currently 60 stars.
+
+.sh
+3.16.1 SELECT Parameters
+
+ The parameters which control the SELECT task are the following:
+
+.ks
+.nf
+Positional or Query Parameters:
+
+ input_group filename
+ output_group filename
+ min_group integer
+ max_group integer
+.fi
+.ke
+
+.le 4
+.ls 16 input_group
+The input group file which is to be searched for groups within the
+limits specified by min_group and max_group.
+.le
+.ls output_group
+The output group file which will consist of groups between 'min_group'
+and 'max_group' in size.
+.le
+.ls min_group
+The minimum group size to be extracted from the input group file.
+.le
+.ls max_group
+The maximum group size to be extracted from the input group file.
+.le
+.le
+.sh
+3.17 The SKY Task
+
+.sh
+3.18 The SORT Task
+
+.sh
+3.19 The SUBSTAR Task
+
+ The SUBSTAR command takes the point-spread function for an image
+frame and a file containing the x,y coordinates and apparent magnitudes
+for a group of stars, usually an output file from one of the photometry
+routines, shifts and scales the PSF function
+according to each position and magnitude, and then subtracts it from the
+original image frame.
+
+.sh
+3.19.1 SUBSTAR Parameters
+
+ The parameters for SUBSTAR control the input and output from this
+task.
+
+.ks
+.nf
+Positional or Query Parameters:
+
+ psf_file filename
+ phot_file filename
+ input_image filename
+ output_image filename
+.fi
+.ke
+
+.ks
+.nf
+Hidden parameters:
+
+ verbose boolean false
+.fi
+.ke
+
+.ls 4
+.ls 16 psf_file
+The name of the file containing the PSF which is to be used as the template
+in the star subtraction.
+.le
+.ls phot_file
+The file containing the photometry results for the stars which are to be
+subtracted from the input image.
+.le
+.ls input_image
+The name of the input image file from which the stars are to be subtracted.
+.le
+.ls output_image
+The name of the output image file which will be a copy of the input frame
+except for the subtracted stars.
+.le
+.ls verbose
+If this parameter is set to true then more information about the progress
+of SUBSTAR is output.
+.le
+.le
+.sh
+4.0 Example
+
+.endhelp