DATE-OBS has unexpected meaning in SharpCapture
Posted: Mon Nov 04, 2019 8:56 am
Hi!
As indicated by the comment inserted in the fits header (done with a ZWO ASI CMOS camera)
DATE-OBS= '2019-10-30T20:07:22.3712127' / System Clock:Frame Received
SharpCapture uses the time it receives an image for the DATE-OBS field. AFAIK this is unexpected behavior, the DATE-OBS field is intended to hold the START of the exposure, e.g. see https://www.plate-archive.org/wiki/inde ... der_format , or https://diffractionlimited.com/help/max ... itions.htm (but see [*] below)
I know that the time the frame was received is actually the only thing SharpCapture can know for sure (modulo read-out delay, transfer delay), for a CMOS camera the better approximation for DATE-OBS should be the time the frame was received minus the exposure time.
This does matter for time series photometry work, occultation timing, exo-planet transit timing etc where second-accuracy is needed, and as analysis software will interpret DATE-OBS as the start of the exposure, you run a risk of getting sub-optimal results unless you are actively (manually, script) correcting the timestamps, which is tedious.
So I would suggest that either DATE-OBS should use [time-frame-was-recived minus exposure-time] , plus a comment saying that this is actually just that, time -received minus exposure time.
What do you think?
EHEA
[*] I'll admit that strictly speaking, this is not actually non-standard.
The FITS standard states: http://archive.stsci.edu/fits/fits_stan ... 0000000000
"The value of DATE-OBS shall be assumed to refer to the start of an observation, unless another interpretation is clearly explained in the comment field".
While the comment does explain SharpCapture's meaning of DATE-OBS in a human readable form, none of the tools I know will be able to make sense of it, and they will just treat this time the way they expect it, as exposure start time. Some plate solving software I used will even re-generate the FITS header and replace the original comment from SharpCapure with it's own like:
DATE-OBS= '2019-10-30T19:17:44.133' / [ISO 8601] UTC date/time of exposure start
(w/o doing any adjustments
).
As indicated by the comment inserted in the fits header (done with a ZWO ASI CMOS camera)
DATE-OBS= '2019-10-30T20:07:22.3712127' / System Clock:Frame Received
SharpCapture uses the time it receives an image for the DATE-OBS field. AFAIK this is unexpected behavior, the DATE-OBS field is intended to hold the START of the exposure, e.g. see https://www.plate-archive.org/wiki/inde ... der_format , or https://diffractionlimited.com/help/max ... itions.htm (but see [*] below)
I know that the time the frame was received is actually the only thing SharpCapture can know for sure (modulo read-out delay, transfer delay), for a CMOS camera the better approximation for DATE-OBS should be the time the frame was received minus the exposure time.
This does matter for time series photometry work, occultation timing, exo-planet transit timing etc where second-accuracy is needed, and as analysis software will interpret DATE-OBS as the start of the exposure, you run a risk of getting sub-optimal results unless you are actively (manually, script) correcting the timestamps, which is tedious.
So I would suggest that either DATE-OBS should use [time-frame-was-recived minus exposure-time] , plus a comment saying that this is actually just that, time -received minus exposure time.
What do you think?
EHEA
[*] I'll admit that strictly speaking, this is not actually non-standard.
The FITS standard states: http://archive.stsci.edu/fits/fits_stan ... 0000000000
"The value of DATE-OBS shall be assumed to refer to the start of an observation, unless another interpretation is clearly explained in the comment field".
While the comment does explain SharpCapture's meaning of DATE-OBS in a human readable form, none of the tools I know will be able to make sense of it, and they will just treat this time the way they expect it, as exposure start time. Some plate solving software I used will even re-generate the FITS header and replace the original comment from SharpCapure with it's own like:
DATE-OBS= '2019-10-30T19:17:44.133' / [ISO 8601] UTC date/time of exposure start
(w/o doing any adjustments
).