Information on NOT FITS files
The NOT FITS headers and file formats are currently being reviewed.
It is foreseen that a small subset of the new keywords described in
the draft FITS requirements
will be implemented in the NOT FITS files in autumn 2004.
Known features, funnies and fixes
We are aware of several short-comings of the NOT FITS headers. Below
you find a list of some notable deviations from standard FITS usage,
and some tips to overcome these. Several of these can be remedied by
applying not2mef to data files. Not2mef can
also be applied retroactively to data from previous periods.
Please share your experiences and
tricks regarding NOT FITS headers and data reduction packages (stp [at]
not.iac.es), so all NOT users can benefit from them.
NOTCAM WCS headers
| Description | NOTCAM WCS
headers are incomplete for NOTCAM data taken in ramp-sampling
(non-destructive readout) mode. This is an error in the image
acquisition software. | | Affects | IRAF imcombine | |
Work-around | Remove the offending WCS headers (CRPIX1,
CRPIX2, CRVAL1, CRVAL2, CDELT1, CDELT2) from the FITS
file. |
|
Fixed by not2mef? | YES. Not2mef processes ramp-sampling data (wich are currently being
recorded in a data cube) into multi-extension
FITS (MEF) and incorporates nominal WCS information. |
Use of DATE-OBS and TM_START keywords
| Description |
Currently, DATE-OBS records the date at the end of
detector read-out in stead of at the start of the integration.
The comment field is empty (for FIES and MOSCA), or misleading ('Start
of observation' for STANCAM, 'End of observation' for ALFOSC).
TM_START correctly records UT at the start of the integration. For exposures
that start integration before midnight, and finish read-out after
midnight, DATE-OBS lists the date after midnight.
For NOTCAM TM_START does not correspond to the the starting time of the exposure itself, but to the reset frame readout preceding the actual exposure.
|
| Affects | | | Work-around
| |
|
Fixed by not2mef? | Yes. Not2mef merges the information from
DATE-OBS and TM_START into the DATE-OBS keyword in ISO8601
yyy-mm-ddThh:mm:ss.s format recording the start of the integration.
A new keyword DATE-AVG has been introduced for the midpoint of the integration. |
Free-format character strings (ao TCSTGT, most string-valued MOSCA, NOTCAM, FIES keywords, FILTER keyword value for STANCAM)
| Description | The position of
the opening quote of several string keyword values is different
from column nr 11. The use of free format character strings is compliant
with the FITS standard, but not recommended. | | Affects | IRAF hselect |
| Work-arounds | - Do the selection
outside IRAF.
- Change
the format with hfix.
|
|
Fixed by not2mef? | Yes. Not2mef records all
FITS headers in the recommended formats. |
MOSCA data format
| Description | The data from
all 4 MOSCA detectors are combined into a single FITS image. The
exposed pixel data from the different detectors in the resulting image
are separated by a few lines of pre- and overscan. This offset does
not correspond to the physical gap between the detectors in the array.
Because of insufficient memory in the detector controller, the pre-
and overscan are limited to a few lines. | | Affects | | | Work-around | |
|
Fixed by not2mef? | Yes. Not2mef copies the single image data into MEF format (image
data for each detector are recorded in a separate FITS extension) and
adds several FITS headers, including nominal WCS information. |
NAXIS3 incorrectly present for MOSCA and FIES
| Description | MOSCA and FIES
data have conflictive mandatory headers regarding the dimensions
of the data array. Data are recorded in a 2-dimensional array
(NAXES equals 2), yet NAXIS3 is also specified. This is an error
in the image acquisition software. | | Affects | | | Work-around | Ignore NAXIS3 for MOSCA and
FIES data |
|
Fixed by not2mef? | Yes. |
DETNAME and PIXSIZE values incorrect for ALFOSC fast-photometry after 2003-09-24
| Description | The values for
DETNAME and PIXSIZE in ALFOSC fast-photometry headers still reflect
the previous ALFOSC detector (CCD7) which was retired in September 2003. The
fast-photometry mode of ALFOSC is using non-standard image acquisition
software. It is not known when this will be updated. |
| Affects | | | Work-around | See detector information |
|
Fixed by not2mef? | No. |
CDELT2 missing for ALFOSC fast photometry
| Description | The WCS keyword CDELT2 (increment in second axis) is not recorded in ALFOSC fast photometry mode, while CDELT1 (increment in first axis) is. The
fast-photometry mode of ALFOSC is using non-standard image acquisition
software. The corrected version of this software is due to be implemented in December 2005.
| | Affects | IRAF imcombine | |
Work-around | For unbinned data, add CDELT2 = 1 to the FITS
header (for binned data CDELT2 = binning-factor). |
|
Fixed by not2mef? | No. |
Use of character " (double-quote, ASCII code 34) in APERTUR keyword values
| Description | The value field
of the APERTUR keyword may contain a " (double-quote, ASCII code
34) meaning arcsec. This character is part of the standard ASCII
character set, and therefore its use within a string valued
keyword is allowed by the FITS standard. (bug reported to
IRAF) | | Affects | IRAF imgets | |
Work-around |
|
Fixed by not2mef? | No. |
Use of character # (cross hatch, ASCII code 35) in AFILTER, BFILTER, FILTER, GRISM keyword values
| Description | The value field
of the AFILTER, BFILTER, FILTER and GRISM keywords may contain a #
(cross hatch, ASCII code 35) meaning number. This character
is part of the standard ASCII character set, and therefore its use
within a string valued keyword is allowed by the FITS standard.
| | Affects |
| | Work-around |
|
Fixed by not2mef? | No. |
|
