[daip] new data weights from FITLD

[Daniel Lebach]

Hello,

This letter is a follow-up inquiry to a letter written very recently
by Michael Bietenholz, who is a colloaborator of mine.

As one who is perennially behind in processing data, I only very
recently enountered for the first time a UV file of VLBI data from
FITLD in which the data weights are not the same as the "tape
weights."  I write to inquire if you can direct me to a document that
describes how the new data weights are defined.

The geodetic-style of processing I employ includes use of my own
task, SN2HF (a derivative of CL2HF), which has relied on tape weights
being available both to determine effective scan durations and to
compute SNRs (SNR being directly proportional to the square root of
the number of bits correlated).  Needless to say, I am going to have
to update SN2HF to be able use these new data weights instead of tape
weights and so must understand how the new weights are defined.
Also, is there any kind of "tag" in a FITS UV file that will tell
you which weights (tape weight or new-style weights) are used, or
do you just have to look at the weights themselves and infer based
on their magnitudes?

Sincerely,
Dan Lebach

P.S. Another issue raised by somebody (Craig Walker?  Michael
Bietenholz?) is the WTTHRESH parameter of FITLD: is a WTTHRESH of 1.0
still guaranteed to unweight all data (or, correspondingly, if you
want, e.g., the tape weight threshhold to be 80%, should you enter
something other than 0.8 with the new data weight scheme)?  In fact
this question more or less reflects the same concern I have with
SN2HF.

--------------------------------------------------------------------------
| Daniel E. Lebach         ! Harvard-Smithsonian Center for Astrophysics |
| dlebach at cfa.harvard.edu  ! 60 Garden Street, MS 42                     |
| Voice: (617) 495-7150    ! Cambridge, MA  02138                        |
| Fax:   (617) 495-7345    ! USA                                         |
--------------------------------------------------------------------------