[alma-config] BOUNCE alma-config at majordomo.cv.nrao.edu: Non-member submission from ["David Woody" <dwoody at caltech.edu>]

[Min Yun]

Due to the pickiness of the mail distributor, this e-mail from Dave
Woody did not reach everyone.  Here it is.   -- Min

------------- Begin Forwarded Message -------------

Received: from cv3.cv.nrao.edu (cv3.cv.nrao.edu [192.33.115.2]) by 
zia.aoc.NRAO.EDU (8.9.3/8.9.3) with ESMTP id PAA28058 for <myun at aoc.nrao.edu>; 
Mon, 7 Feb 2000 15:42:11 -0700 (MST)
Received: from kochab.cv.nrao.edu (kochab.cv.nrao.edu [192.33.115.108]) by 
cv3.cv.nrao.edu (8.9.3/8.9.3/CV-SOL-3.0) with ESMTP id RAA24852 for 
<myun at nrao.edu>; Mon, 7 Feb 2000 17:42:09 -0500 (EST)
Received: (from majordom at localhost) by kochab.cv.nrao.edu 
(8.9.3/8.9.3/CV-SOL-3.0) id RAA03326; Mon, 7 Feb 2000 17:42:08 -0500 (EST)
From: owner-alma-config at kochab.cv.nrao.edu
Date: Mon, 7 Feb 2000 17:42:08 -0500 (EST)
Message-Id: <200002072242.RAA03326 at kochab.cv.nrao.edu>
To: owner-alma-config at kochab.cv.nrao.edu
Subject: BOUNCE alma-config at majordomo.cv.nrao.edu:    Non-member submission from 
["David Woody" <dwoody at caltech.edu>]   
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Length: 9154

>From myun at nrao.edu  Mon Feb  7 17:42:05 2000
Received: from zia.aoc.NRAO.EDU (zia.aoc.nrao.edu [146.88.1.4])
	by kochab.cv.nrao.edu (8.9.3/8.9.3/CV-SOL-3.0) with ESMTP id RAA03321
	for <alma-config at kochab.cv.nrao.edu>; Mon, 7 Feb 2000 17:42:04 -0500 
(EST)
Received: from ovro.ovro.caltech.edu (ovro.ovro.caltech.edu [192.100.16.2])
	by zia.aoc.NRAO.EDU (8.9.3/8.9.3) with ESMTP id PAA28053
	for <alma-config at zia.aoc.NRAO.EDU>; Mon, 7 Feb 2000 15:42:00 -0700 (MST)
Received: from woodyspc (new [192.100.16.90])
	by ovro.ovro.caltech.edu (8.9.3/8.9.1) with SMTP id OAA24487
	for <alma-config at zia.aoc.NRAO.EDU>; Mon, 7 Feb 2000 14:41:59 -0800 (PST)
Message-ID: <006201bf71bd$4f04b750$5a1064c0 at ovro.caltech.edu>
From: "David Woody" <dwoody at caltech.edu>
To: "ALMA config" <alma-config at zia.aoc.NRAO.EDU>
Subject: Fw: [alma-config] Re: UV sampling metric
Date: Mon, 7 Feb 2000 14:47:27 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.00.2919.6600
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600

> Hi Min
> Following is my spin on the metric issue.  I wrote it while the
discussions
> were going on and so does not address the detailed math.  The discussion
> is great.  I wish I had more time to contribute.
> Cheers
> David
> **********************************************************
>
> ALMA configuration issues
>
> The configuration issue is obviously complex and we can only arrive
> at a technical consensus if we come up with a well defined and
> mutually agreeable metric.
>
> _________________________________________
> Image metrics vs. UV metrics
>
> Test images and image quality figures of merit is one
> approach.  But that approach inherently limits us to using the
> existing imaging algorithms and our expectations of what
> future ALMA images will look like.  The algorithms evolve/improve
> continuously and it is unlikely that they will be the algorithms
> of choice for an instrument that gets first light in ~10 years
> and will operate for at least another 20-30 years.
> Mosaic imaging will be an important part of this
> instrument and the mosaic algorithms will probably change
> dramatically in the next 10 years.
>
> (Some of the imaging done now is limited by our knowledge
> of and ability to deal with the detailed beam patterns
> for each antenna.  This is particularly important for
> mosaicing or even single position imaging of complex fields.
> ALMA will be significantly better in this respect than any
> existing array.  Hopefully we will develop algorithms and
> have the computing power to use all of the detailed beam
> information from the individual antennas.)
>
> The simulation approach using imaging algorithms
> assumes that there is information in the UV-plane that is not
> required to make the kind of images we want.  If this is the case,
> then you can always improve your sensitivity, resolution,
> fidelity or dynamic range, etc. by building configurations that
> don't measure the "unneeded" UV data.  This can be used to
> optimize instruments designed for a specific observation
> or set of observations.
>
> But ALMA is supposed to cover
> essentially all of millimeter and submillimeter science
> now and for the next 30 years.  It would be
> a mistake to try and guess what the "unneeded" UV data is
> based upon our current images and algorithms.  It is almost
> axiomatic that any configuration optimized for these images
> and algorithms will not be optimized for entirely new images
> or algorithms.
>
> This is why we should look to more fundamental metrics, such as
> UV coverage or completeness.  Ed Fomalont has made some good
> suggestions on how to improve this type of metric by
> including smoothness and lumpiness criteria, etc.
> I believe this is fundamentally a more sound approach.
>
> At the same time we should use the image simulations to verify
> that the various configurations do not jeopardize the kinds
> science that we know will be done.  We should not make decisions
> based upon differences of a few percent in the image quality
> metrics but make sure that we are not losing a lot more.
>
> ******************************************
> One way to bring these two views together is to look at the FT
> of the test images and see what the amplitude in the UV plane
> looks like.  This will show us what kind of UV data is needed
> independent of the imaging algorithm.  Then for a particular
> configuration we can look at the UV data that is missed and
> see what features in the image will be missed.  It may indeed be
> that there will be image information that the astronomers
> don't want or need.
>
> It would be great if we could show that configurations that
> have good UV metric values also have good image metric values
> and visa versa.  If this were true, then the search for
> the "best configurations" would be much faster and
> straight forward and would not depend upon current algorithms.
>
> This could lead to an entirely different way of picking the
> configurations instead of simply having scaled versions giving
> different resolutions.
> Each configuration would optimize different aspects of the
> UV coverage and hence the kind of science, such as
> 1)  maximize short baselines (surface brightness sensitivity),
>  i.e. close packed array.
> 2)  maximize resolution with "complete" coverage.  This also
> gives you uniformity and smoothness.
> i.e. 2-3 km approximate rings,see memo 270.
> 3)  best tapered UV coverage with complete coverage, i.e.
> 1 km donut or spiral.
> 4)  maximum resolution, i.e. largest possible circle.
> 5)  uniform coverage at different resolutions
> 6)  tapered coverage at different resolutions
> The latter two configuration types would fill in if we thought
> the first 4 types left large holes in the angular resolution
> coverage.
> Of course there is also the complexity of the source declination,
> hour angle and earth rotation to deal with.
> *********************************************************
> ________________________________________________________
> memo 270
>
> The point of memo 270 was to show that you
> can achieve nearly complete coverage using a few hours of earth
> rotation at an interestingly high resolution, i.e. 2 km baselines.
> If the maximum baselines had been much shorter, then it would
> not have been very useful or interesting.
>
> Even longer baselines
> and more uniform coverage should be possible using
> Keto type rings.  Pushing the
> maximum baseline with complete coverage automatically moves you
> toward the better uniformity and smoothness in the UV coverage
> that Ed Fomalont proposes as metrics.
> This is only one configuration
> and all of the bigger and smaller configurations can be picked
> based upon other considerations, such as nicely taper
> UV coverage etc.  But we could confidently tell the astronomers
> and NSF and ESO that we will produce images with >~2 km baselines
> with "no missing information".  We can even mosaic large fields
> at this resolution, again with "no missing information".
>
> I have not done any image simulations on this array.
> So this work applies only to UV coverage and simulated
> imaging type quality factors were not addressed.
>
> In memo 247 Leonia Kogan showed that a uniform close spacing
> circle has large sidelobes as would be expected from a very
> regular array pattern.  The simulated images are not as good
> as for other configurations.
> But it is not clear that this in not a
> deficiency of the algorithms applied to this kind of data.
> The data is complete and simply
> binning the UV data into a 2-D array with proper weighting
> and taking the FT might produce a better image.
> At any rate, a uniform circular array has
> very strong UV weighting at the longest and shortest
> baselines, i.e. is far from uniform UV coverage.  Other
> configurations can have much better UV metric values.
>
> _____________________________________________________________
> Other considerations:
>
> Snap shot images will be a very common
> mode and the instantaneous UV coverage or image from a configuration
> is important, but all of the configurations should usefully
> exploit earth rotation to improve the coverage and images.
>
> Clearly logistics and cost will be important, but at this
> early stage we should concentrate on the scientific goals
> for ALMA.  Once we have done this, we can then see what
> compromises are required for it to be affordable and
> to improve the operation logistics.  I don't believe there
> is a hard budget break down at this point and any
> compromises should be made in terms of optimizing
> the science output for a given cost across the whole project.
> *****************************************************
> | David Woody
> | Assistant Director of Instrumentation
> | Owens Valley Radio Observatory
> | California Institute of Technology
> | 100 Leighton Lane (PO Box 968)
> | Big Pine, CA 93513
> | 760-938-2075ext111, FAX 760-938-2075
> | dwoody at caltech.edu
> *****************************************************
>


------------- End Forwarded Message -------------