NAME

deosc - remove ``oscillations'' from event lists


SYNOPSIS

deosc options [input file]


OPTIONS

deosc uses long options. Options may be abbreviated, and the ``='' character shown below in the option templates is optional. A command line argument which begins with the @ character is taken to be the name of a file which contains white-space delimited options.

Input Data Specification

--input=file
The input events FITS file. The file may also be specified as the last argument on the command line.

--extname=string
The FITS extension name of the HDU to be read from the input file. It defaults to events

--xcol=string --ycol=string --tcol=string
The names of the event position and time columns to use. These default to x, y, and time, respectively.

--gti=file
A FITS file containing a GTI extension. This should be a CFITSIO extended filename which explicitly specifies the GTI HDU.

--scale=float
A factor by which the input event positions will be multiplied (for units conversion). It defaults to 1.

--sky_xcol=string, --sky_ycol=string
The names of the event positions in sky coordinates. Used only if --skyfilt is set. These default to x and y, respectively.

--units=string
The position units (after scaling by --scale). For plotting purposes. Defaults to pixels.

Output Options

--tag=string
The prefix used for the output files. This parameter is required.

--device=PGPLOT device
The output plotting device. It defaults to /xs, which plots to the screen. Useful other ones are /cps and /vcps, which are, respectively, landscape and portrait color PostScript output.

Hardcopy will be left in files with extensions of .ps for PostScript, .png for PNG, etc. Output to the screen via the /xs device will be sent to mutiple windows. The default /xs device number to start at is 5; this may be changed by specifying the device number directly: 6/xs.

--plot, --noplot
Do (or don't) generate the plots. Defaults to --plot.

--title=title
The plot title. If not specified it is gleamed from the input file's header.

--wtitle, --nowtitle
Do (don't) write a title on thep lots. Default is to write the title.

--write, --nowrite
Write out the modified data to the FITS file tag_fix.fits. Please note that the output positions are scaled by the --scale parameter! Defaults to --write.

--wcsub, --nowcsub
If true, the original and ``fixed'' event positions as well as the centroid fits written to the FITS file will have the center of the image subtracted from them. It defaults to --wcsub.

See Output for more information.

--width=float
Specify the (approximate) width of the resultant plots, in inches.

--ocols
A comma separated list of the columns to write out. The available columns are:
     fitx fity
     inclip
     time time0
     x x0
     y y0

The default list is

     time
     x x0
     y y0

--verbose
Be noisy.

Image Centering parameters

The following parameters apply to centering in the coordinates given by the --xcol and --ycol parameters:

--force_ctr
If specified, the point specified by the --xc and --yc options will be used as the center, rather than being determined by an iterative sigma clipping method. The --clip_r option must be set if --force_ctr is used.

--dtol=float
Absolute convergence tolerance for the iterative centering algorithm. It defaults to 0.01.

--iclip=float
The initial centering clipping radius, in scaled position units. It defaults to 1000.

--iter=integer
The number of iterations the centering routine should perform. It defaults to 10.

--nsigma=float
The clipping radius, in units of the standard deviation of the event distribution about the center, used during the iterative centering. It defaults to 3.

--xc=float
If set, this specifies an initial value (in scaled units) for the X coordinate of the center of the object. See --scale.

--yc=float
If set, this specifies an initial value (in scaled units) for the Y coordinate of the center of the object. See --scale.

--pixcent
If specified, choose the brightest pixel in the image as the initial center. This option optionally takes an integer argument specifying the image size in pixels (the image is square). This defaults to 256.

The following parameters apply to centering in the coordinates given by the --sky_xcol and --sky_ycol parameters, if --sky_filt is used. See Sky Filtering.

--sky_filt
If specified, an initial center is determined based upon the coordinates given by --sky_xcol and --sky_ycol. The events within a circular aperture of radius given be either --sky_clip_r or --sky_clip_sigma are then used to determine the center in --xcol and --ycol coordinates.

--sky_force
If specified, the point specified by the --sky_xc and --sky_yc options will be used as the center, rather than being determined by an iterative sigma clipping method. The --sky_clip_r option must be set if --sky_force is used.

--sky_dtol=float
Absolute convergence tolerance for the iterative centering algorithm. It defaults to 0.01.

--sky_clip=float
The initial centering clipping radius, in scaled position units. It defaults to 1000.

--sky_iter=integer
The number of iterations the centering routine should perform. It defaults to 10.

--sky_nsigma=float
The clipping radius, in units of the standard deviation of the event distribution about the center, used during the iterative centering. It defaults to 3.

--sky_xc=float
If set, this specifies an initial value for the --sky_xcol coordinate of the center of the object.

--sky_yc=float
If set, this specifies an initial value for the --sky_ycol coordinate of the center of the object.

--sky_pixcent
If specified, choose the brightest pixel in the image as the initial center. This option optionally takes an integer argument specifying the image size in pixels (the image is square). This defaults to 256.

--sky_clip_sigma=float
The radius of the circular region around the object used to select events for filtering in coordinates given by --xcol and --ycol, in units of the calculated standard deviation of the event distribution about the center. See also --sky_clip_r.

--sky_clip_r=float
The radius of the circular region around the object used to select events for filtering in coordinates given by --xcol and --ycol, See also --sky_clip_sigma. This must be specified if --sky_force is used.

Analysis Options

--clip_sigma=float
The radius of the circular region around the object used for the analysis, in units of the calculated standard deviation of the event distribution about the center. See also --clip_r.

--clip_r=float
The radius of the circular region around the object used for the analysis, in scaled units. See also --clip_sigma. This must be specified if --force_ctr is used.

--seglen=float
The length (in seconds) of segments used in fitting the event positions as a function of time. Defaults to 100s. See DESCRIPTION.

Miscellaneous Options

--help
Output abbreviated usage information and exit.

--usage
Output extensive usage information and exit.

--version
Output version information and exit.


DESCRIPTION

deosc removes time dependent ``oscillations'' from an input event list by modelling the motion of an object's centroid and subtracting it from the events.

The correction is determined by fitting the gross motion of events in a circular region centered around the object, as a function of time.

The input events must be in a binary table extension of a FITS file. CFITSIO is used to extract the data, so the CFITSIO extended filename syntax may be used to filter the data. The event positions may be scaled (to convert to seconds of arc, for example).

Determining the Object Center

If --force_ctr is specified, the source center is provided by the --xc and --yc options.

If --force_ctr is not specified, the center is determined from the data. If the event coordinates are in a reference frame which is dithered (such as Chandra chip coordinates), multiple objects may overlap in space (but not time), and it may be difficult to accurately determine the object center. In this case, it may be possible to select events associated with the object based on their dither-corrected coordinates (such as Chandra sky coordinates). This is accomplished by what is termed here sky filtering. The center of the object in sky coordinates is determined, a region around that center is excised, and the contained events are used to derive the center in dithered coordinates.

Sky Filtering

In the case, as described above, that the analysis is to be performed in dithered coordinates, but those result in spatial confusion of sources, a subset of source events may be selected by centering and extracting events in dither-corrected, or sky coordinates.

To accomplish this, specify --sky_filt, and set the following parameters. --sky_force --sky_dtol, --sky_clip, --sky_iter, --sky_nsigma, --sky_pixcent, --sky_xc, --sky_yc, --sky_xcol, --sky_ycol.

These are akin to the centering parameters described in the following section. The size of the region to use as the source events is determined by the --sky_clip_r or --sky_clip_sigma parameters.

Centering

If --force_ctr is not specified, the location and scale of the object is determined by iteratively centering on the event coordinates as given by the --xcol and --ycol parameters. The initial center may be specified in one of the following methods:

If none of these methods is specified, the average event position is used.

An initial clip is performed about the initial center, with the clipping radius provided by the --iclip parameter. Events outside of a clipping radius are discarded, and the average center is redetermined.

This step is repeated until either the center hasn't changed, or the absolute change in the center is less than the tolerance specified by the --dtol option. After the initial clip, the clipping radius is a multiple of the standard deviation of the events; the --nsigma option specifies the mutiplication factor.

Motion correction

The radius of the region used to fit the time-varying motion is determined either from the standard deviation of the event distribution about the center, if deosc determined the center and --clip_sigma is specified, or may be given directly via the --clip_r option.

Within this region the events are divided (in time) into subsets composed of three segments, each with a length specified by --seglen. The subsets are staggered such that each overlaps the preceding by two segments. The one-dimensional differences (along each coordinate) in event positions from the determined object centroid are fit as a function of time by polynomials of maximum order 10. The actual number of orders various by subset, and is that which produces the best fit. The fit is done in a least-squares sense, so the resultant polynomials track the center of the event ``cloud'' (PSF).

The resultant fits to the events' motion in the inner segment of each subset is then used to correct the positions of all of the events which fall within that time slice. No quantitative effort is made to ensure the continuity of the fits across subsets. The overlapping of the subsets and the fitting across neighboring segments seems, empirically, to constrain the fits so that the polynomials in neighboring segments are reasonably continuous.

The correction of events which occur outside the circular analysis region may introduce errors in the positions of events whose time motion is not correlated with that of events in the analysis region, but this is unimportant if the former are used only for background measurements.


OUTPUT

Several files are output upon successful completion:


VERSION

$Revision: 1.24 $


AUTHOR

Diab Jerius (djerius@cfa.harvard.edu)