iraf.net

>From Brian.Taylor@lowell.edu Fri Jan 23 15:43:10 2004
Subject: - Never mind
From: "Brian W. Taylor" <Brian.Taylor@lowell.edu>
To: valdes@noao.edu
Date: Fri, 23 Jan 2004 15:42:48 -0700Hi Frank, Nevermind, I was able to get it straightened out. It was
due to a conflict with the HEADER and the instrument data file. Thanks!
Brian>From Phil.Massey@lowell.edu Mon Jan 26 14:20:43 2004
Date: Mon, 26 Jan 2004 14:20:09 -0700 (MST)
From: Phil Massey <Phil.Massey@lowell.edu>
To: fvaldes@noao.edu
Subject: CCDSEC and DATASEC
Hi, Frank---I admit that the behavior of ccdproc with our headers remains mysterious
despite your useful note. If I do a setinstrument direct, and then run
ccdproc with just overscan and trim set, I get:cc> ccdproc
List of CCD images to correct (a*.fits): sue.fits
ERROR: Size of DATASEC and CCDSEC do not agreeBut our headers only contain CCDSEC, not DATASEC. According to
"ccdgeometry", if DATASEC is absent, only entire image is assumed to
be the data. But, actually what seems to happen is that we get the
above error message.thanks,
phil>From valdes@noao.edu Fri Jan 30 10:18:15 2004
Date: Fri, 30 Jan 2004 10:18:02 -0700 (MST)
From: Frank Valdes <valdes@noao.edu>
To: Phil.Massey@lowell.edu
Subject: Re: CCDSEC and DATASECHi Phil,Sorry about not responding sooner. I had a high priority project to complete
by the end of the month. The best way to respond to questions about CCDPROCs
complaints about keywords is to get header listings. So if you are
still having trouble understanding the DATASEC/CCDSEC mismatch error please
send me an example header.Yours,
Frank>From Phil.Massey@lowell.edu Fri Jan 30 10:29:47 2004
Date: Fri, 30 Jan 2004 10:29:04 -0700 (MST)
From: Phil Massey <Phil.Massey@lowell.edu>
To: valdes@noao.edu
Subject: Re: CCDSEC and DATASECFrank,Here's the header listing. Note that there is no DATASEC entry, which
"ccdgeometry" implies should be OK, but it isn't. I've tried modifying
CCDSEC to conform with the entire data section [1:1024,1:1024] but it
didn't seem to help. In other words, if DATASEC is missing it seems to
fail.thanks!
---philLine storage mode, physdim [1065,1024], length of user area 1661 s.u.
Created Tue 00:00:00 01-Jan-1980, Last modified Tue 00:00:00 01-Jan-1980
Pixel file "040127.010.fits" [ok]
BZERO = 32768 / zero point
BSCALE = 1 / data scaled by value
BUNIT = 'ADU ' / pixel units(ADU,electrons)
LOISVERS= '1.3.3.3 ' / LOIS Version
LCCDMOD = 'loral ' / LOIS ccd module
LTELMOD = 'perkins ' / LOIS telescope module
LINSTMOD= 'lowell10' / LOIS instrument module
OBSERVAT= 'Lowell Observatory' / observatory
ELEVATIO= '2200m ' / altitude
LATITUDE= '35.09683' / latitude, degrees
LONGITUD= '-111.53592' / east longitude, degrees
OBSERVER= 'sled jet' / observer(s)
OBSAFFIL= 'CSUSB/NASA JSC & LCO' / observer(s) affiliation
DETECTOR= 'Loral SN1259' / CCD Detector Name
CCDMODE = 'Single ' / CCD exposure mode
GAIN = 3.100000E+00 / Gain of array in Electrons per ADU
RDNOISE = 1.700000E+01 / Readnoise of array in Electrons
CCDSUM = '2 2 ' / CCD Binning
ORIGSEC = '[1:1065,1:1024]' / Full Binned Frame
CCDSEC = '[16:1032,2:1023]' / Best Image Section
1.700000E+01 / Readnoise of array in Electrons
CCDSUM = '2 2 ' / CCD Binning
ORIGSEC = '[1:1065,1:1024]' / Full Binned Frame
CCDSEC = '[16:1032,2:1023]' / Best Image Section
CCDTEMP = -1.276700E+02 / CCD Temp in Deg C
SETTEMP = -1.250268E+02 / CCD Temp Set Value in Deg C
PIXTIME = 2.000000E+00 / readout time in microsecs per pixel
TELSCOP = 'PERKINS ' / Telescope name
ST = '02:07:50' / sideral time
RA = '03:45:03.3' / right ascension(hh:mm:ss)
DEC = '+39:43:57' / declination (dd:mm:ss)
EQUINOX = 2.000000E+03 / equinox of RA and DEC
EPOCH = 2.000000E+03 / same as EQUINOX(for back compat.)
AIRMASS = 1.060000E+00 / airmass
TELFOCUS= -2.900000E+01 / telescope focus position
TELFILT = 5 / telescope filter position
INSTRUM = 'Lowell Filter Instrument Module' / instrument name
FILTERS = 5 / Filter Poisiton
FILTNAME= 'I ' / Filter Name
CCDFNAME= '040127.010.fits' / Orig. CCD disk file name
EXPTIME = 3.000000E+00 / actual integration time, seconds
DARKTIME= 3.000000E+00 / total elapsed time, seconds
IMAGETYP= 'FLAT ' / object, flat, bias, etc.
DATE-OBS= '2004-01-27' / UT date (yyyy-mm-dd) of observation
UT = ' 01:11:17' / universal time (start of exposure)
AT ' / object, flat, bias, etc.
DATE-OBS= '2004-01-27' / UT date (yyyy-mm-dd) of observation
UT = ' 01:11:17' / universal time (start of exposure)>From valdes@noao.edu Fri Jan 30 11:40:57 2004
Date: Fri, 30 Jan 2004 11:40:37 -0700 (MST)
From: Frank Valdes <valdes@noao.edu>
To: Phil.Massey@lowell.edu
Subject: Re: CCDSEC and DATASECHi Phil,What you have isImage size = [1065,1024]
CCDSEC = '[16:1032,2:1023]' / Best Image Section
DATASEC not specified, defaults to image size = [1:1065,1:1024]
TRIMSEC not specified, defaults to image size = [1:1065,1:1024]
BIASSEC not specified, defaults to image size = [1:1065,1:1024]So DATASEC is 1065x1024 while CCDSEC is 1017x1022 which disagree. The
simplest thing you can do is delete CCDSEC and then in CCDPROC you
would explicitly specify BIASSEC and TRIMSEC. The main use of CCDSEC,
other than documenting the readout region, is to allow subsection
readouts to be matched against calibrations that might have been taken
in full format. If you are not doing any of that you can just delete
the keyword.Frank
>From Phil.Massey@lowell.edu Fri Jan 30 11:59:44 2004
Date: Fri, 30 Jan 2004 11:59:01 -0700 (MST)
From: Phil Massey <Phil.Massey@lowell.edu>
To: valdes@noao.edu
Subject: Re: CCDSEC and DATASECHi, Frank---OK, that's great, and explains EVERYTHING. Our goal here is to fix the
headers (and in particular to get the headers right for some new
instruments coming on line.---phil

>From Phil.Massey@lowell.edu Fri Jan 30 13:46:49 2004
Date: Fri, 30 Jan 2004 13:46:11 -0700 (MST)
From: Phil Massey <Phil.Massey@lowell.edu>
To: valdes@noao.edu
Subject: Re: CCDSEC and DATASEC
Frank, one followup question: do you still rely on CCDSUM to specify the
binning? In other words, if we had two CCD frames, one unbinned (the flat,
say), and one binned (a program frame), what would ccdproc expect to see?
Say that the unbinned frame was 1:1024,1:1024 with the usual 32 columns as
overscan. Would it be:Flat:
CCDSEC = [1:1024,1:1024]
DATASEC= [1:1024,1:1024]
CCDSUM= ' 1 1'object frame:
CCDSEC = [1:512,1:512]
DATASEC= [1:512,1:512]
CCDSUM= ' 2 2'and it would work OK?thanks,
phil>From valdes Fri Jan 30 14:00:13 2004
Date: Fri, 30 Jan 2004 14:00:01 -0700 (MST)
From: Frank Valdes <valdes>
To: Phil.Massey@lowell.edu
Subject: Re: CCDSEC and DATASECYes there is some ambiguity about what to do with on chip summing. It requires
some rigorous definition of what the various sections mean in this case.
For MSCRED I did this and CCDSEC is to be specified in unbinned coordinates
while DATASEC is specified in binned coordinates (i.e. image pixels).But in the older CCDRED this was never clearly defined. The result
is that when summing one uses a fake CCDSEC; i.e. one that considers
the binned pixels to be the CCD pixels. So what you show is correct
for imred.ccdproc but would not be correct for mscred.ccdproc.To avoid possible confusion by my explanation, what you show is correct.Frank