The idea here is to give you a two or three sentence explanation of a concept, and (possibly) a link to where you can find more information.
.. glossary::
amplifier
Each CCD in the HSC camera is read-out through 4 separate
amplifiers (amps). The pixel regions read through a given
amplifier are along the columns; thus each 2048 x 4096 CCD has
four 512 x 4096 amps. Each amp behaves slightly differently
electronically, so each of the four CCD regions corresponding
to the amps has a slightly different gain, and non-linearity.
The :ref:`HSC camera layout <hsc_layout>` shows the locations
of amp1 for each CCD in the camera.
aperture flux
aperture photometry
The HSC pipeline measures source flux with various algorithms
(aperture, PSF, cmodel). The 'aperture flux' has traditionally
referred to a straight sum of the counts in all pixels within an
'aperture' or specified (usually circular) region around the
source. What the HSC pipeline uses is conceptually exactly the
same thing, except that it uses a Sinc interpolation algorithm
(developed originally in SDSS) to handle the partial pixels
correctly. See :term:`Sinc photometry <sinc flux>` for details.
background matching
One difficulty encountered in creating a coadd stack of images is
that the sky background present in each image can be very
different (observed on different nights, with different moon
position/phase, etc). The solution has traditionally been to
model the background (usually by smoothing the image with a large
smoothing kernel) and subtract it prior to making the coadd.
However, a better approach may be to choose one image as a
'reference', and instead model the difference between its
background and that of the other input images. In this way, the
input image backgrounds are matched to the reference and stacked.
The result is that the coadd has a non-zero background (that of
the reference image), but its signal-to-noise ratio is higher and
it can therefore be measured and subtracted more accurately than
it could have been prior to co-addition.
brighter-fatter
brighter-fatter effect
To gain sensitivity to red photos, the CCDs used in the HSC camera
(and many modern astronomical camera) are thicker than the
previous generation of devices (200 um versus ~15-25 um). The
pixels have physical dimensions of 15x15 um, and so each can be
thought of as having the aspect ratio of a sky-scraper.
Photo-electrons are released in the higher floors of the silicon
sky scraper and are pulled by an electric field down to the
PN-junction in the basement where they're stored until read-out.
However, as more photo-electrons accumulate in the basement, their
presence tends to deflect some of the newly arriving
photo-electrons into the neighbouring pixels. The concequence of
this is that **brighter** stars have a systematically wider
(i.e. **'fatter'**) point spread function (PSF).
Butler
Rather than having different modules of the pipeline keep track of
where and how they read/write their inputs/outputs, a software
tool called the 'butler' does this. If you go to an expensive
restaurant, the 'valet' will park your car for you. You don't
need to know where the garage is when you arrive, and you don't
need to remember where you parked when you leave. This is what the
butler does for the input/output operations of the pipeline.
Rather than hard-coding the paths, filenames, and loading/writing
syntax for various data inputs and outputs throughout the pipeline
code; you simply make a call to the butler to 'get' the thing you
want for a specific dataId (frame, CCD, etc). For example,
loading a bias image looks like this:
``biasImg = butler.get('bias', dataId)``
CAS
Catalog Archive Server
This term is inherited from SDSS and refers to the online
database system which is used by the community as one of the
main ways to obtain SDSS data. The data available through
this system are the output measurements of the sources (right
ascension, declination, ugriz magnitudes, etc), but not the
images (see :term:`DAS` for that).
.. todo:: perhaps a link here?
ccd
A charge-couple device, of course, but also used to refer to data
from a CCD in a raw data image or in single-frame data products.
See also :term:`DataId`.
cmodel
todo
.. todo:: ask Jim.
CoaddPsf
todo
.. todo:: ask Jim.
DAS
Data Archive Server
This term is inherited from the SDSS and refers to the online data
repository where data products such as images can be obtained.
The pipeline outputs (RA, Dec, magnitudes, etc) are provided by
the :term:`CAS`.
dataId
Individual exposures are refered to either as 'visits' or
'frames', and their sub-components are the CCDs in the camera
(note that LSST refers to these as 'sensors'). However, when
making a coadd, the celestial sphere is broken into a set of fixed
regions called 'tracts', which are similar in size to the field of
view of the HSC camera. The tracts are composed of sub-regions
called 'patches'. Each patch is about the size of a CCD. Thus,
'visit' and 'CCD' are used to refer to raw data or single-frame
data products, while 'tract' and 'patch' refer to coadd data. A
dataId also includes such things as field name (field), date of
observation (dateObs), and filter.
See also :ref:`DataId <back_dataId>`
deblend
Sources which are detected in the pipeline are often found to be a
group of multiple blended/overlapping sources. In order to
measure each of the contributing sources separately, the detected
source (called a 'parent') must be 'deblended' into its
'children'. The algorithm use is decribed in
.. todo:: add link to explanation of deblend algorithm.
deep survey
todo
double-Gaussian
The point spread function of a star is quite similar to a 2D
Gaussian, but has too much flux present in the extended 'wings'.
However, while one Gaussian is a poor model, two Gaussians does
quite a good job. One Gaussian models the center of the PSF,
while the second Gaussian (typically 2x the width and 0.1x
amplitude) models the wings. A double-Gaussian PSF is available
in the pipeline software, but is not used as a PSF for production
reruns.
differencing
todo
doxygen
Doxygen is a code documentation system used by the software group.
The system uses specially formatted comments in the code to
construct a web-based navigable tool which is useful for
developers. The HSC doxygen is served `here
<http://hsca.ipmu.jp/doxygen/>`_.
EUPS
EUPS is the package management system used by the software group.
It's used to install and generally keep track of different
versions of the software pipeline modules. For specific details,
see the :ref:`EUPS page <back_eups>`
extendedness (classification.extendedness)
This is an output value associated with each source measured by
the pipeline. It's stored as a float, but is currently used as a
flag for star/galaxy separation (0 = star, 1 = galaxy ... a galaxy
is more 'extended' than a star).
flag
In any measurement that the pipeline makes, any concerns
associated with the pixels or the measurement will be recorded in
the catalog outputs in parameters with names including the word
'flag'. Examples include ``flags_pixel_edge``,
``flags_pixel_interpolated_any`` ... you can guess what these
mean. The full list is included in the `data products document
<http://hsca.ipmu.jp/hscsoft/datainfo.php>`_.
footprint
Within the software group, the region of pixels occupied by a
source (which we want to measure) is called the source's
'footprint'. The pixels within the footprint are used for the
measurement, the ones outside are not.
forced measurement
In our stacked images, we're able to detect faint sources which
would be below our 5-sigma thresholds in any of the input images,
or in the coadds from different filters. However, once we know
that a source is detected in e.g. a deep i-band stack, we can then
measure it at the location we expect it to be in another image
where it wasn't detected. This is called a 'forced measurement'.
frame
A full exposure including all CCDs. It's assigned a number by the
observatory (called an EXP-ID). The software group tends to use
the LSST term 'visit' to mean the same thing.
FRAMEID
The Subaru name used for a single CCD exposure. It uses
the form ``<4-char><8-digit>``. See :ref:`Data Format <data_format>` for
details.
healpix
There are various ways you can break up the celestial sphere into
discrete regions (called tesselation). HealPix is a popular one
in the astronomy community.
..todo:: We support this, but I don't know of anywhere where we're currently using it.
HSM
Hirata-Seljak-Mandelbaum
This refers to a collection of shape measurement algorithms
coded-up, bundled together, and made public by Chris Hirata,
Uros Seljak, and Rachel Mandelbaum. The package includes 'KSB'
(HSM_KSB), 'regaussianization' (HSM_REGAUSS), 'Bernstein-Jarvis'
(HSM_BJ), 'linear' (HSM_LINEAR), and a shapelet-based algorithm
(HSC_SHAPELET). It is enabled by default in the HSC pipeline.
Kron flux
todo
.. todo:: write this.
KSB
todo
.. todo:: Do we need this? Out of scope for this glossary?
mosaic
Mosaic is the name of the HSC software module which performs a
photometric uber-calibration, in which the photometry measured in
different visits is tied together into the same self-consistent
photometric system.
multifit
todo
.. todo:: ask jim.
multishapelet
todo
.. todo:: ask jim.
object
This is a common word, but has a somewhat specific meaning in the
software group. It refers to a celestial object whose properties
we'd like to measure. It should not be confused with a 'source',
which is a specific exposure instance of an object. For example,
a star is an 'object', but two exposures of it will yield two
'sources'.
patch
See :term:`DataId`.
peak
During :term:`deblending <deblend>`, individual components are
identified in the parent source's :term:`footprint`. The
highest pixel in each candidate child is it's peak.
Petrosian flux
todo
.. todo:: ask rhl.
pipeline
The collection of data processing steps which run autonomously to
take the raw input data and produce the final catalog output
measurements.
PSF
point spread function
The response function of an imaging system to a 'point source', or
delta function. This includes the atmosphere plus the telescope
plus the camera. The PSF is a function of position across the
field of an image, and is also varies over time.
PSF flux
PSF photometry
While aperture photometry measures the sum of the integrated flux
within a synthetic aperture around a source; PSF photometry
measures the *weighted* sum of the flux, where the weight function
is the local PSF at the sources position in the image. If the
source is itself a point source (i.e. a star), then the
measurement is optimal.
PSF-Ex
A PSF model library developed by Emmanuel Bertin. PSF-Ex is used
for PSF flux measurement in the HSC pipeline.
raft
The LSST camera (*not* HSC) is subdivided into 21 square
platforms, with 9 CCDs mounted 3x3 on each one (total 189
CCDs). The 21 square platforms are called 'rafts'. The HSC
camera is not structured this way, but you may occassionally hear
the term as the pipeline code is shared with the LSST project.
rerun
The term ``rerun`` originated in SDSS. It simply refers to a
single processing run, performed with a specified version of the
reduction code, and with a specific set of configuration
parameters. The implication is that within a given 'rerun', the
data have been handled in a homogeneous way.
schema
(w.r.t. database) The schema of a database is its structure.
It refers to the coded blueprint which describes how the data
are to be stored with respected to one another. Which fields
will appear in which tables, and what types of data they will
contain are described in a database's schema.
Note also, that the HSC database system uses PostgreSQL, and the
term 'schema' has a specific meaning in the context of postgreSQL.
Separate logical databases within a single PostgreSQL database are
called 'schema'.
sensor
See :term:`DataId`.
sinc flux
sinc photometry
todo
skymap
todo
Sloan swindle
todo
source
todo
SSP
See Strategic Survey Proposal
stack
(w.r.t. the data reduction pipeline) A slang term for the
complete set of software packages which make up the pipeline
code.
stack
(w.r.t. image coaddition) A synonym for coadd.
Strategic Survey Proposal (SSP)
todo
TAN-SIP
todo
Task
Each step in the pipeline processing is contained within a
software class called a 'Task'.
tract
See :term:`DataId`.
uber-calibration
Uber-calibration was originally developed in SDSS to tie all
observations onto a single consistent photometric system. The
method relies on repeated observations of the same objects in
multiple exposures. The calibration terms can then be adjusted to
allow measurements in the different exposures to be compared
meaningfully. The HSC uber-calibration process is called 'mosaic'.
.. todo:: put a ref to Nikhil's paper.
ultra-deep survey
todo
visit
See :term:`DataId`.
warp
In order to produce a stack, the input images must all be
resampled onto a common pixel grid. The process is referred to as
warping.
WCS
World Coordinate System
todo
wide survey
todo