... work,^2.1

Or even if it does.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... program:^2.2

If the program wasn't installed in /usr/local/bin or similar, you may have to type the full path to the binary; from the distribution toplevel directory type: src/gcx

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... button.^2.3

The image will pan only up to the point where it's edge is at the edge of the window.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... drag^2.4

move the mouse while holding the left button pressed

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... frame.^2.5

The fields that are always required, like naxis, SIMPLE, BITPIX, etc are not displayed.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... advance.^2.6

The symbols used to depict various star types can be set by the user, so their appearance can vary. These are the default shapes.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... required.^3.1

This function uses the zcat utility, which must be installed on the system.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Header.^3.2

Note that the required FITS fields like SIMPLE, BITPIX, NAXIS, NAXIS2, BSCALE, BZERO are not shown by this operation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... button.^3.3

The image will pan only up to the point where it's edge is at the edge of the window.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... WCS.^4.1

See the ``World Coordinates'' section below.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... routine.^4.2

The conversion routines are located in recipy.c; The function called by the import function is convert_catalog.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... GSC-ACT^4.3

GSC-ACT is a recalibrated version of the GSC. While the random astrometric errors are about the same, the systematic errors have been reduced considerably.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... file:^4.4

This make-shift arrangement is likely to change in future versions.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... epoch.^5.1

Whenever the epoch of some coordintes is not specified, GCX assumes J2000.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... fields.^5.2

This is the same system used by the popular wcstools package.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... order:^5.3

The actual names of any of the fields can be changed in the options page. The names below are the defaults.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... fails,^5.4

The author would very much like to receive any frames for which the algorithm fails.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... follows:^6.1

Neglecting non-linear terms in the CCD response.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... of:^6.2

Neglecting the noise the bias frames add to the dark current estimate.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... camera^6.3

For the purpose of this discussion, the telescope illumination non-uniformity is folded into the camera response; the sensitivity we estimate will correct both the camera and the telescope's response nonuniformity.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... obtain:^6.4

We can in principle replace the master flat frame with this normalised frame, by which we can divide the data frames directly. However, a normalised frame doesn't represent well in the very common 16-bit integer format. So we instead choose to keep the master flat frame scaled to it's original scale, and take care of the normalisation in the flat-field division routine.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... it.^6.5

All processing is done on a copy of the frame held in memory--the original frames are not altered in any way.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... name^6.6

If the output file name field is left blank, the result of the stacking operation will not be saved. It will however display in the main window, and can be saved from there.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...6.6^6.7

The Output file / dir name field must contain a file name if the result of the current reduction set is a single frame (stacking is enabled) or a directory name if the result of the current reduction set consists of several frames (stacking is disabled).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... again,^6.8

Note that once a reduction operation has been performed on a image frame in the list, it will not be performed again unless the image frame is reloaded--which discards any changes to it. If we want to apply different sets of reduction operations to some frames, like creating frames with different scales, the frames have to be reloaded (selected in the Image Files list and Reload clicked) before changes in parameters (like the multiplication factor or bias file name) can take effect.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... names,^6.9

Or just the dark file name if we work without biases.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... images).^6.10

The matching algorithm does provide rotation and scale information, so this option is planned to be added in a future version.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... window.^6.11

The stars will only show if an image frame is currently being displayed.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... frame.^6.12

On fields with a large number of stars, it is sometimes possible that the default settings result in a bad match. If this is the case (a good indication of this is an unusually hish shift displayed in the Run report) increase the Wcs Fitting Options/Minimum number of pairs value by a few units.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... photometry).^7.1

Without taking the color of the stars into account. Computing and applying color transformation coefficients requires using the information from multiple frames, taken with different filters. The output of the aperture photometry operation can be fed into the multi-frame reduction routine which takes care of that.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... image.^7.2

Choosing too large an aperture has two drawbacks: it is more likely for the aperture to include unwanted stars; and the signal to noise ratio is degraded, especially for faint stars. A common way around this is to emply variable-size apertures, and correct the data for the amount of light that falls outside the aperture. While this method can give good results, it requires that the star images be well-sampled and uniform is shape across the frame. It also needs a good model for the light sensitivity distribution inside a pixel. Neither of the above conditions are met by common amateur setups, so great care is required when working with variable apertures.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...mode^7.3

The most frequent value

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... targets^7.4

This will still not be the ``true'' magnitude of the star, because we haven't taken the colors of the standard and target into account. It is however the best we can do from a single frame. See Chapter 8 for fitting color transformation coefficients and other multi-frame reduction operations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... error.^7.5

When using multiple standard stars, we have the ``luxury'' of using the actual spread of their magnitudes to calculate the zeropoint error, as opposed to the estimated errors.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... stars.^7.6

A recipy file can be created ``on the fly'' by marking stars and then editing them to enter the standard magnitudes and types (standard and target). A valid WCS is required before being able to edit frame stars. In a pinch, using Wcs/Force Validate will make the program think it has one. Of course, all the coordinates will be off in the report file. It is much better to create a proper recipy file first (Section 7.7).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... this:^7.7

This way we don't need to keep CCD-reduced version of the data frame, but rather work directly on the original frames.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... format,^7.8

The output format can be set from the command line using the -S option.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... band,^7.9

The band names are case-insensitive.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... are:^8.1

These assume that a linear color transformation coefficient is enough to transform the data. While this is a largely used assumption, it is by no means guaranteed for any data set. One should carefully check the data to determine if a linear transformation is appropiate

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... eliminated.^8.2

When we fit frame zero points without taking the transformation coefficients, the zero points are ``exact'' for stars having the color index equal to the mean color index of the standard stars in each frame. The adjusted zeropoints are ``exact'' for stars with a color index of zero.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... extinction^8.3

The amount of light lost when passing through the atmosphere.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... following:^8.4

R.H. Hardie, 1962, Photoelectric Reductions, Chapter 8 of Astronomical Techniques, W.A. Hiltner (Ed), Stars and Stellar Systems, II (University of Chicago Press: Chicago), pp178-208.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... fields.^8.5

The standard fields can contain differential photometry targets.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... extinction.^8.6

The program allows a small amount of extrapolation to take place.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...ap:robust,^8.7

A robust regression, rather than averaging algorithm is used, which uses the same outlier down-weighting scheme.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... is:^8.8

It is common to include additional coefficients in the model, such as ambient or sensor temperature or time. But they are generally only effective in photometric conditions, where the extinction varies smoothly over long intervals of time.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... repeteadly.^8.9

If we have a large numbers of files to add, it's probably easier if they are concatenated before loading (using cat for instance).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... set.^8.10

The default setting will work if gnuplot is installed in the command path.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... directly.^8.11

The data frame is a simple ascii table with some gnuplot commands at the top; it can easily be imported in other plotting utilities or spreadsheet programs if desired.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... downward.^8.12

If the data was reduced with GCX the saturated stars would normally be marked as such and excluded for the fit. If down-going branches appear, the Saturation limit option's value should be decreased.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... more).^8.13

The plot routine clips standard errors at $\pm30$, and residuals at $\pm2$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... zero.^8.14

The most dangerous deviant points in this case are not the ones with large standard errors (which are easily detected), but the ones right near the turning point of the graph. Being bright stars, they will have small estimated errors, and can bias the solution significantly. In this particular case it didn't happen because of the large number of ``normal'' stars.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... index).^8.15

It is important at this step to examine the data carefully and check if a simple linear transformation coefficient will remove any color trend in the data. Some data sets may show a curved dependence (for which a polynomial transformation would be better), while other can show turning points. In those cases, the linear transformation no longer holds.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... (0.002),^8.16

This error propagates to the final magnitude in proportion with the target star's color index. An coefficient error of 0.002 will contribute a systematic error of 0.004 to a star with $B-V=2.0$, and 0.001 to one with $B-V=0.5$

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... line.^8.17

The zeropoints increase when the transparency improves.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... frames,^8.18

The status of the frames that have had a good extinction fit will end in ``-AV''.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... frames.^8.19

The points on the time graph can appear to wander more than the error bars would indicate. This is because the values there have not been corrected for airmass, and the airmass does not correlate with time.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... is:^A.1

This assumes that the method used for sky estimation has a statistical efficiency close to the mean, which isn't generally the case. Perhaps this should be taken into account, at least for methods whose efficiency is well known, like the median.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... function.^B.1

See Peter B. Stetson, The Techiques of Least Squares and Stellar Photometry with CCDs at http://nedwww.ipac.caltech.edu/level5/Stetson/Stetson_contents.html.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.