1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
|
.help findgain Nov01 noao.obsutil
.ih
NAME
findgain -- calculate the gain and readout noise of a CCD
.ih
SYNOPSIS
FINDGAIN uses Janesick's method for determining the gain and read noise
in a CCD from a pair of dome flat exposures and a pair of zero frame
exposures (zero length dark exposures).
.ih
USAGE
findgain flat1 flat2 zero1 zero2
.ih
PARAMETERS
.ls flat1, flat2
First and second dome flats.
.le
.ls zero1, zero2
First and second zero frames (zero length dark exposures).
.le
.ls section = ""
The selected image section for the statistics. This should be chosen
to exclude bad columns or rows, cosmic rays and other blemishes, and
the overscan region. The flat field iillumination should be constant
over this section.
.le
.ls center = "mean"
The statistical measure of central tendency that is used to estimate
the data level of each image. This can have the values: \fBmean\fR,
\fBmidpt\fR, or \fBmode\fR. These are calculated using the same
algorithm as the IMSTATISTICS task.
.le
.ls nclip = 3
Number of sigma clipping iterations. If the value is zero then no clipping
is performed.
.le
.ls lsigma = 4, usigma = 4
Lower and upper sigma clipping factors used with the mean value and
standard deviation to eliminate cosmic rays.
Since \fBfindgain\fR is sensitive to the statistics of the data the
clipping factors should be symmetric (the same both above and below the
mean) and should not bias the standard deviation. Thus the values should
not be made smaller than around 4 sigma otherwise the gain and readnoise
estimates will be affected.
.le
.ls binwidth = 0.1
The bin width of the histogram (in sigma) that is used to estimate the
\fBmidpt\fR or \fBmode\fR of the data section in each image.
The default case of center=\fBmean\fR does not use this parameter.
.le
.ls verbose = yes
Verbose output?
.le
.ih
DESCRIPTION
FINDGAIN uses Janesick's method for determining the gain and read noise
in a CCD from a pair of dome flat exposures and a pair of zero frame
exposures (zero length dark exposures).
The task requires that the flats and zeros be unprocessed and uncoadded so
that the noise characteristics of the data are preserved. Note, however,
that the frames may be bias subtracted if the average of many zero frames
is used, and that the overscan region may be removed prior to using this
task.
Bad pixels should be eliminated to avoid affecting the statistics.
This can be done with sigma clipping and/or an image section.
The sigma clipping should not significantly affect the assumed gaussian
distribution while eliminating outlyers due to cosmic rays and
unmasked bad pixels. This means that clipping factors should be
symmetric and should have values four or more sigma from the mean.
.ih
ALGORITHM
The formulae used by the task are:
.nf
flatdif = flat1 - flat2
zerodif = zero1 - zero2
gain = ((mean(flat1) + mean(flat2)) - (mean(zero1) + mean(zero2))) /
((sigma(flatdif))**2 - (sigma(zerodif))**2 )
readnoise = gain * sigma(zerodif) / sqrt(2)
.fi
where the gain is given in electrons per ADU and the readnoise in
electrons. Pairs of each type of comparison frame are used to reduce
the effects of gain variations from pixel to pixel. The derivation
follows from the definition of the gain (N(e) = gain * N(ADU)) and from
simple error propagation. Also note that the measured variance
(sigma**2) is related to the exposure level and read-noise variance
(sigma(readout)**2) as follows:
.nf
variance(e) = N(e) + variance(readout)
.fi
Where N(e) is the number of electrons (above the zero level) in a
given duration exposure.
In our implementation, the \fBmean\fR used in the formula for the gain
may actually be any of the \fBmean\fR, \fBmidpt\fR (an estimate of the
median), or \fBmode\fR as determined by the \fBcenter\fR parameter.
For the \fBmidpt\fR or \fBmode\fR choices only, the value of the
\fBbinwidth\fR parameter determines the bin width (in sigma) of the
histogram that is used in the calculation. \fBFindgain\fR uses the
\fBimstatistics\fR task to compute the statistics.
.ih
EXAMPLES
To calculate the gain and readnoise within a 100x100 section:
.nf
ms> findgain flat1 flat2 zero1 zero2 section="[271:370,361:460]"
.fi
To calculate the gain and readnoise using the mode to estimate the data
level for each image section:
.nf
ms> findgain.section="[271:370,361:460]"
ms> findgain flat1 flat2 zero1 zero2 center=mode
.fi
The effects of cosmic rays can be seen in the following example using
artificial noise created with the \fBartdata.mknoise\fR package. The
images have a gain of 5 and a readnoise of 10 with 100 cosmic rays added
over the 512x512 images. The zero level images have means of zero and the
flat field images have means of 1000. The first execution uses the default
clipping and the second turns off the clipping.
.nf
cl> findgain flat1 flat2 zero1 zero2
FINDGAIN:
center = mean, binwidth = 0.1
nclip = 3, lclip = 4., uclip = 4.
Flats = flat1 & flat2
Zeros = zero1 & zero2
Gain = 5.01 electrons per ADU
Read noise = 10.00 electrons
cl> findgain flat1 flat2 zero1 zero2 nclip=0
FINDGAIN:
center = mean, binwidth = 0.1
nclip = 0, lclip = 4., uclip = 4.
Flats = flat1 & flat2
Zeros = zero1 & zero2
Gain = 2.86 electrons per ADU
Read noise = 189.5 electrons
.fi
.ih
BUGS
The image headers are not checked to see if the frames have been
processed.
There is no provision for finding the "best" values and their errors
from several flats and zeros.
.ih
REVISIONS
.ls FINDGAIN - V2.12
New task derived from MSCFINDGAIN. This makes use of the new clipping
feature in IMSTATISTICS.
.le
.ih
SEE ALSO
imstatistics
.endhelp
|
