[daip] help with UVLSF, please

[Elias Brinks]

Hi Eric,

Do you mind if I pick your brain for a moment? The bigwigs in
charge have been less then helpful to LITTLE THINGS, the
large VLA project we got oodles of time on, with respect to
the aliasing problem and basically left us to our own devices.
I guess the excuse is that Michael Rupen is on the project so
he should take care of solving the problem.

All of our targets will be observed at either 1.5 MHz or 0.78 MHz
in straight 2AD mode (so no overlapping IFs). We should have
a good number of line-free channels on either side of the line
emission (with good I mean more than sufficient for an ORDER=1
UVLSF). The targets have virtually no continuum associated with
them, so I'd only expect the extragalactic 20cm sky to come in,
with the odd annoying bright background source somewhere in
the field.

This means that fitting an ORDER=4 continuum makes things
worse (as illustrated by the plots in yesterday's e-mail). I could
do a UVLSF with ORDER=0 based on the unaffected side of the
bandpass. This is not ideal as some continuum sources have a
spectral index which warrants them being removed with ORDER=1.

Another approach is to use UVLSF with order=1 on all
baselines except EVLA-EVLA and then do a UVLIN with
ORDER=0 on the unaffected side on the EVLA-EVLA.
Problem is, one isn't subtracting out the aliased continuum
which enters on EVLA-EVLA baselines.

I guess my real question is, how does this aliased emission get
imaged? Is the phase scrambled to such an extent that
it would just add "noise", albeit with some unpredictable
structure? So, should I just "live with it"?

Would you have any other suggestion on how to deal with it?

Thanks very much for any suggestions you might have.

Cheers,

Elias



On 11 Dec 2007, at 22:12, Eric Greisen wrote:

> Without even looking at the plots, I think I can answer to some
> extent.  Higher order fits are really not justified by the noise in
> single records of vis data.  In particular, the high order fits are
> not constrained by data in those channels excluded from the fit.  If
> there are too few channels for the S/N and the higher order then
> things will get out of hand in the unconstrained regions of the
> spectrum.  In particular, you will see bad results in scalar averaging
> in which the fact that the phase also got out of hand in those regions
> does not protect you from bad signals.  vector averaging with phases
> wandering wildly will reduce the large amplitudes that appear.
>
> UVLSF with high order will only work if there is really a significant
> continuum signal and you can use e.g at least half the channels in the
> fit.  In particular, the lower numbered channels need to be line
> free.
>
> In other words, this change to UVLSF is not magic although it works
> well in a limited set of circumstances.
>
> Eric Greisen