iraf.net

Hello.
I have a few questions and comments about the new version of
IRAF recently released.1) There is a bug in splot in the function command when operations are
performed between two spectra with different binning. Apparently there
is no way to reset the way the rebinning is done prior to an operation,
and the flux is conserved per bin unit. A simple test to demonstrate
this is to take a spectrum, rebin it (into say twice as many pieces)
using dispcor and setting "conserve flux" to "no". This will keep the
flux levels the same, but simple change the sampling. Now, dividing
the two using splot will not yield a spectrum with average 1, but with
some other value (but strangely enough, not 2). I think just changing
the flux conservation parameter in splot is enough, but this is something
not contained in the parameter list.2) There is another bug, or at least something inconvenient in the
extraction of 2 dim spectra to 1 dim spectra. When my 2D images are
reduced to 1D using apsum or apall, the new file type with bands, etc.,
is used. That is fine, but I can no longer simply run the longslit
packages fluxcalib and extinction on the 1D images because the
dispersion information is completely rewritten into a new format,
and the parameters like crval1 are removed that the longlist packages
look for.3) Obviously, because of #2 above, the standard procedure of longslit
spectroscopic reductions outlined in the noao manual for the older
versions (by Ed Anderson) is not valid. Is there an update for this
version available now that you could send me?Thanks,Mike Brotherton
University of Texas at Austin
Astronomy Dept.

Hi Mike,You are correct that a missing functionality in SPLOT, SCOMBINE, and
SARITH is the ability to select whether flux over pixels should be
conserved when spectra are resampled to common pixels. I did the
test you described an it did work properly. I also did not get an
exact factor of 2 but this was because you have to be careful about
how the endpoints enter. For example a 512 point spectrum will have
it's dispersion halved by specifying an output of 1023 points and
not 1024. The additional bands in APEXTRACT is controlled by the "extras" parameter
in APSUM/APALL. If you don't want this stuff set the value to no.There is a logical problem in that longslit.extinction and
longslit.fluxcalib are intended for long slit spectra and not 1D spectra.
If the cookbook recommends using these tasks on the output of APSUM then I
think this is an error even though it may have previously worked
correctly. The tasks to use with 1D spectra are onedspec.calibrate. This
task does understand and properly operate on all data produced by APSUM
including the one with the "extras" bands. There is no update on the
longslit cookbook which is supposed to deal with the part of distortion
correcting and calibrating 2D longslit spectra. There have been no
significant changes in this area. Any discussion of extracting 1D spectra
from 2D longslit spectra should be treated as illustrations. Note I have
not gone and looked up the cookbook but I think I have answered your
immediate question. Let us know if you have further questions.Yours,
Frank Valdes