[daip] Alignment problem.
R. Craig Walker
cwalker at nrao.edu
Tue May 1 02:29:41 EDT 2007
I am trying to make my M87 polarization images and did something that is
probably not too common, but I don't think too unreasonable. I used
multi-resolution clean for the IPOL on a fairly large, complicated source.
But for Q etc, I used only one resolution since the polarization is
barely detected, if at all. Then I did the usual combinations with COMB,
and tried to plot the results with PCNTR, showing both the IPOL and
polarized images. It dies with a complaint about the images being
insufficiently aligned. I am using a 7 mas RASHIFT. For the IPOL image,
the reference pixel is in the center:
AIPS 1: Image=3C274 (MA) Filename=BW088A3C274 .ICL001. 59
AIPS 1: Telescope=VLBA Receiver=VLBA
AIPS 1: Observer=BW088 User #= 501
AIPS 1: Observ. date=27-JAN-2007 Map date=30-APR-2007
AIPS 1: Minimum=-1.23253872E-03 Maximum= 6.06720865E-01 JY/BEAM
AIPS 1: ----------------------------------------------------------------
AIPS 1: Type Pixels Coord value at Pixel Coord incr Rotat
AIPS 1: RA---SIN 1024 12 30 49.423 512.00 -.000035 .00
AIPS 1: DEC--SIN 1024 12 23 28.044 513.00 .000035 .00
AIPS 1: FREQ 1 4.3134928E+10 1.00 2.6000000E+07 .00
AIPS 1: STOKES 1 1.0000000E+00 1.00 1.0000000E+00 .00
AIPS 1: ----------------------------------------------------------------
AIPS 1: Coordinate equinox 2000.00
AIPS 1: Map type=NORMAL Number of iterations= 1511
AIPS 1: Conv size= .00035 X .00020 Position angle= .00
AIPS 1: Observed RA 12 30 49.423 DEC 12 23 28.04
AIPS 1: Phase shifted in X -.007 in Y .000
AIPS 1: Rest freq .000 Vel type: OPTICAL wrt YOU
AIPS 1: Alt ref. value 0.00000E+00 wrt pixel .00
Here is the position from the UV file:
AIPS 1: Image=3C274 (UV) Filename=BW088A3C274 .CALIB . 59
AIPS 1: Telescope=VLBA Receiver=VLBA
AIPS 1: Observer=BW088 User #= 501
AIPS 1: Observ. date=27-JAN-2007 Map date=30-APR-2007
AIPS 1: # visibilities 424868 Sort order TB
AIPS 1: Rand axes: UU-L-SIN VV-L-SIN WW-L-SIN TIME1 BASELINE
AIPS 1: INTTIM GATEID CORR-ID WEIGHT SCALE
AIPS 1: ----------------------------------------------------------------
AIPS 1: Type Pixels Coord value at Pixel Coord incr Rotat
AIPS 1: COMPLEX 1 1.0000000E+00 1.00 1.0000000E+00 .00
AIPS 1: STOKES 4 -1.0000000E+00 1.00 -1.0000000E+00 .00
AIPS 1: FREQ 1 4.3126928E+10 1.00 1.3000000E+07 .00
AIPS 1: IF 2 1.0000000E+00 1.00 1.0000000E+00 .00
AIPS 1: RA 1 12 30 49.423 1.00 3600.000 .00
AIPS 1: DEC 1 12 23 28.044 1.00 3600.000 .00
AIPS 1: ----------------------------------------------------------------
AIPS 1: Coordinate equinox 2000.00
AIPS 1: Rest freq .000 Vel type: OPTICAL wrt YOU
AIPS 1: Alt ref. value 0.00000E+00 wrt pixel .00
The position is the same as in the UV file and points to the center.
To get the right position, a program would need to pay attention also to
the phase shifted line.
On the QPOL image, we get:
AIPS 1: Image=3C274 (MA) Filename=BW088A3C274 .QCL001.1059
AIPS 1: Telescope=VLBA Receiver=VLBA
AIPS 1: Observer=BW088 User #= 501
AIPS 1: Observ. date=27-JAN-2007 Map date=30-APR-2007
AIPS 1: Minimum=-6.82381773E-03 Maximum= 5.12634544E-03 JY/BEAM
AIPS 1: ----------------------------------------------------------------
AIPS 1: Type Pixels Coord value at Pixel Coord incr Rotat
AIPS 1: RA---SIN 1024 12 30 49.423 312.00 -.000035 .00
AIPS 1: DEC--SIN 1024 12 23 28.044 513.00 .000035 .00
AIPS 1: FREQ 1 4.3134928E+10 1.00 2.6000000E+07 .00
AIPS 1: STOKES 1 2.0000000E+00 1.00 1.0000000E+00 .00
AIPS 1: ----------------------------------------------------------------
AIPS 1: Coordinate equinox 2000.00
AIPS 1: Map type=NORMAL Number of iterations= 50
AIPS 1: Conv size= .00035 X .00020 Position angle= .00
AIPS 1: Phase shifted in X -.007 in Y .000
AIPS 1: Rest freq .000 Vel type: OPTICAL wrt YOU
AIPS 1: Alt ref. value 0.00000E+00 wrt pixel .00
The POLC and POLA image headers look the same as the QPOL image.
This header has the same coordinates as IPOL, and the same phase shifted
indication, but a very different reference pixel - the actual source
location in the image.
Interestingly, TVLAB seems to understand what has happened as loading
with the same blc/trc and asking for LTYPE 3 gives the same results for
both IPOL and QPOL. But it looks like PCNTR doesn't understand. It's not
obvious to me how to tell what happened - is it encoded in keywords
PARANGLE and ZENANGLE. Those seem to be the only things that look
different. But from what I can tell in a quick look at the HELP files,
these aren't an obvious place for an important item like a phase shift.
My polarization images look bad. The highest points are over 10 sigma and
are not on-source. The usual N-S grot for equitorial sources is there,
but what is going on at the source doesn't seem reasonable. Stuff is
spread out, kind of like you might get with phase errors. I don't know if
I just haven't detected the polarized flux or if something went wrong in
calibration. My two calibrators have reasonable looking polarization
starting from the same CL table. A puzzle for tomorrow.
Actually, maybe the apparent header discrepency is simple. The header
positions are rounded to the nearest milli time second (RA) and two
numbers separated by 7mas, appropriately situated, could round to the same
value. TVLAB does print rather more significant figures. Basically the
numbers from IMHEAD are marginal to insufficient for VLBI.
But the question still stands. Why does PCNTR consider two images made
with a tget IMAGR, as far as the image geometry is concerned, to be
insufficiently aligned?
I must be tired (it's after midnigh). The above was with DOALIGN=1.
For DOALIGN=0, PCNTR is happy. But it still seems like the images should
satisfy the DOALIGN=1 criteria given the similarity of the method by which
they were made. Is it subject to roundoff effects trying to relate the
images with different reference pixels?
Cheers,
Craig
---------------------------------------------------------------------
R. Craig Walker Array Operations Center
cwalker at nrao.edu National Radio Astronomy Observatory
Phone 505 835 7247 P. O. Box O
Fax 505 835 7027 Socorro NM 87801 USA
--------------------------------------------------------------web
More information about the Daip
mailing list