- refs and news
- These are the names of images that go into the light curve. Each
"new" is an image in which the supernova light will be
photometrically measured. The same aperture is measured in each
"ref" to figure out how much host galaxy light must be subtracted
from each new in order to figure out the brightness of just the
supernova at the time of the new. Generally, all of your refs
should also be news, so that they will be measured and plotted on the
lightcurve.
You want your refs to be images of decent seeing, and to be deep as
well, because they are subtracted from all the other images. If your
refs are noisy, then your lightcurve will by necessity be noisy. You
also want your refs to have as little supernova signal as possible.
At any time, you can change the refs and news in a lightcurve software
by giving ltcvparent the /change_i switch. This is discussed
elsewhere. (Yell at Rob once the elsewhere is written and force him to
put a link from here to there.)
- A Primary Reference
- This is not to be confused with the refs (above), for
the primary reference does not have to be one of your "refs". It
may well be one of your "news", but again it doesn't have to be. The
primary reference (or "primary ref") is a coordinate reference. All
positions are stored relative to the primary reference. All photometry
is normalized to the primary reference. In other words, output files
which report
a number of counts don't give the number of counts on the new in
question, but
rather the number of counts on the new times the ratio in counts of a
constant object between the primary ref and that new.
You probably want your primary reference to be a deep image with decent
seeing. You find all of your fiducial objects and (probably) your
standard stars on the primary ref with the ltcvparent program,
so you want to stack the odds in your favor by starting with a good
image.
Once a primary reference is defined for a lightcurve, it will forever
more be that primary reference, unless you completely start the
lightcurve over. Even if you change your refs, the primary reference
stays the same.
- The positions of the supernova and its host
- In pixels on the primary ref, of course. Don't muck with RA&Dec,
because we don't know those to the sort of sub-pixel accuracy we
need (especially with frames as warped as BTC frames). It is
originally set based on the positions saved to the candidate netcdf
file by the searchscan software; later, you can manually change
the position of the supernova based on results returned by the Ivan
lightcurve software.
- Fiducial Objects
- Also sometimes called "standard galaxies," although this is a slight
misnomer. This list is composed of most of the other objects within
the same flat region on the primary reference as the supernova.
The centers (peaks, cores, whatever) of the fiducial objects are stored
in the parent file, in pixel positions on the primary reference.
They are used for two purposes. One, measurement of apertures on the
fiducial objects, offset from the center by the same vector
offset of the SN from its host, are used to determine the normalization
ratio between the counts of the a and the counts of the a when
subtracting the ref from the new. Two, the lightcurves of these offset
positions on the fiducial objects are measured in the same manner as
the lightcurve of the supernova. Since the fiducial objects should not
be varying, the lightcurves should be flat within the photometric
errors. If they aren't, we worry, and the software adds an additional
error to reflect the scatter of the fiducial object lightcurves.
- Standard Stars
- Positons (in pixels on the primary reference) of stars on the primary
reference. In addition, the parent file has the flux (in counts) of
the standard stars
within the smae flat region of the primary reference as the
supernova.
Positions of stars all over the whole field.
They should be real point sources, not
galaxies, and isolated (i.e. without any companions too close). They
are used for two purposes. One, the stars which are in the mutual flat
region of a ref and a new are used to measure the point spread function
(psf) of the ref and
the new. The software convolves the sharper image with a Gaussian
until the psfs of the two images match. Then, the psf of the resultant
images is used to determine the size of the aperture which should be
used to measure the supernova and the amount of reference light to
subtract out. (By default, an aperture of radius 1*FWHM is used.)
The second use of the standard stars is to measure the flux ratio
between each new and the primary reference. Once the amount of
supernova light on a new has been measure, before it is reported it is
multiplied by an outlier-clipped variance-weighted average of the
ratios in counts of the standard stars on the primary ref and that new.
(Only
the standard stars in the mutual flat region of the primary ref
and the new are used for this.)
(The standard stars may also be used to compute spatial transforamtions
between images, but I belive that this is not actually the case -- in
reality, full photometry lists are used as usual with the deepsearch
software.)