Expand GUI to h-a Filter name.
(AJR -> NGD)
Data reduction techniques
- BUGS and other problems with the data reduction pipeline! (NGD has updated
this list and corrected most of the bugs) -- only
major problems are listed in this todo file!
- - DONE by ND
- Xstartrails could be placing seeing and weight estimates in the image
header, this could assist in paper-writing, OR (better?) generate
a table DATE, EXP, SEEING
- Discuss a MAJOR PROBLEM: With present strategy (aligning
via star spectra) the alignment can fail when the filter profile
changes, and we think we see this happening from night to night.
It can be corrected for IF
we have a arc+mask and tungsten+mask image for each night. The problem
can be corrected in xstartrails but
an extra routine is required to fine-tune the
alignment of the images automatically: WHO? ND has incorporated
an arbitrary shift into xstartrails so the errors, if present,
can be corrected without degrading the data.
Nigel suggests "dithering" the stack of images (in X only)
to optimise signal at (say) four locations.
- Update website to ensure that certain
tasks (eg arc+mask and tungsten+mask) are done each night.
I.e "Night checklist" required! Make sure one image is braketted
by a stable mask pair
- Check which holes can/must be used in a given mask. NOTE THE FOLLOWING POINT!!
- Use a summed mask for each night to provide a wavelength
solution to Calctrans - this will have higher S/N and
will allow more wavelengths to be used. Use the individual
masks, as before, to align images (AJR has produced stackmasks.cl to do
this) but it does not seem to enhance the
visibility of the spots enough (NGD will experiment and report)
- Possible bug: codes 501728 and 501725 have been used for the same
wavelength. UNTIL FURTHER NOTICE WE IGNORE BLEND ISSUE
AND DESIGNATE ALL 5017 BLENDED LINES 5017 (ABOVE CODES ARE CANCELLED)
- Check vignetting/shutter function using the obs of stars/galactic PNe
and implement the shutter correction procedure
- Other gaps in reduction pipeline?
- Iterate within team
and write up findings on source extraction;
Look for ways to discriminate background galaxies from PNe: line width, etc,
check for known catalogued sources in field (KK,NRN).
Think about PHOTOMETRY and CENTROIDING; is PSF fitting of interest?
- more Monte Carlo image simulations to estimate
completeness limits, contamination fraction, etc;
also differing effects over field
(PSF and vignetting at edges, brightness nearer centre,
star trails obscuring) (NRN)
PSF fit to M31 data? (HRM polynomial solution-> AJR, NRN)
Establish quantitatively what our
rejection limit is: look
for rotation curve in galaxies (OII emission from bkg
galaxy at z=0.35) - resolve doublet?;
objective comparison of object shape with PSF from other nearby objects;
compare predictions from large scale structure simulations
or observations (Steidel) -- compare density, luminosity function, clustering
- derive empirical S/N relation with time, etc. and check predictions
- measure bandpass over the field for different filter angles,
using TUNGSTEN maks: issues are
- filter cutoff for planning observations: use master mask?
- (currently done for B-6°, B-0°;)
- post-hoc analysis of completeness: use master mask?
- photometry: preferably use run masks?
B-6: bandpass is characterised, but not yet checked for changes with time; high-quality arc is available
B-3, etc: do we ever need to use such in-between angles?
B-0: bandpass is characterised, but not yet checked for changes with time; high-quality arc is available
A-0: bandpass needs characterisation; high-quality arc available??
C-0: bandpass needs characterisation; high-quality arc available but out of focus
C-6: bandpass needs characterisation; high-quality arc available but out of focus
D: to be delivered