[ATM] (Long) How good is good enough? (was how much celldeformation is too much?)

["Scott Milligan" <starzkey@charter.net>]

Ah the question that never dies!  R. Feynman has gone on record saying that
"It's the subject that nobody knows anything about that we can all
talk about".  But what about questions that we know something about, but
perhaps not everything?  Then it seems, we argue.

 

After following the thread (loosely) on the subject of how much cell induced
deformation is too much, my interest was alerted when the subject steered
onto the question of whether an observer can distinguish between optics
rated in the laboratory as diffraction limited (as evidenced by a Strehl
ratio evaluation > 0.8), and optics that test measurably better than that
standard.  The subject is interesting (to me) both because of the many
conflicting opinions that have been expressed over the years in attempting
to answer it, and also because the fact when it comes time to discuss the
impact on visual observations, data obtained from anything like a controlled
experiment is conspicuously lacking.

 

In the professional literature, the question has received both theoretical
and empirical attention; however, since professional astronomers are by and
large unconcerned with visual observation of the scene, the most useful
results of their analyses have been directed at answering the question “how
good does my telescope have to be given the fact that I am making a long
exposure image through a randomly varying phase-amplitude screen? (i.e. the
atmosphere). 

 

Having read some of these papers, I have broadly summarized (in my own mind)
the most important results as concluding that under the best average seeing
conditions available on planet earth, limiting resolution increases with
aperture for telescope apertures < about 20 inches, where after the
atmosphere dominates the image forming process, imposing an upper bound on
the resolution that can be achieved.  The assumptions underlying that
summary are that a long exposure image is formed through a “perfect” optical
system, and that modern image enhancing technologies such as adaptive
mirrors are not used.  The arguments followed to arrive at this conclusion
(Fried, 1964), have been extrapolated to stipulate that an acceptable
standard for the figure accuracy of large astronomical optics (i.e. >> 20
inches aperture, and again without relying on adaptive optics) be such that
residual slope errors across a large majority of the optical surface should
be no greater than +/- the ½ the angular extent of a star image as limited
by atmospheric seeing.  If we accept that most amateur astrophotography is
(historically at least) resolution limited by factors other than optical
quality (i.e. atmosphere, detector pixel size), then I think it is
relatively easy to convince ourselves that we do in fact have a lot of
“data” to support the argument that long-exposure imaging is typically
dominated by factors other than optical quality (i.e. wavefront error) as
measured on the bench.  Witness the many fine photos taken using commercial
SCT’s, which are known (on average) to be detectably “imperfect” when
subjected to any of the optical tests used by ATM’s.  Perhaps the only
exception I can think of off-hand is that the presence of excessive
secondary spectrum when exposing bright objects IS usually detectable and
detrimental to the results, but most instruments so afflicted will have
apertures considerably less than 20 inches, and so should be expected to
demonstrate resolution that is less likely to be limited by atmospheric
seeing.

 

That the situation might possibly be quite different for visual observing
should, perhaps, not come as a surprise, for now we are adding a “smart”
detector into the system, i.e. the human eye-brain.  Factors such as
residual optical figure errors, self-weight deflection or cell-induced
mounting strains remain essentially constant during the time an observation
is made.  Thermal effects and atmospheric “seeing”, in contrast, are only
meaningfully described in terms of their statistically derived average
impact during the observation interval.  Hence, attempting to mathematically
“add” these two error classes may not be very meaningful unless the
observation interval is >> longer than the minimum length interval required
to collect a statistically valid sampling of the random variable’s behavior.
The commonly accepted re-fresh rate for the human eye is about 1/30 second;
if, on a steady night, the average period of atmosphere induced phase
fluctuations is longer than this, then I conclude it is reasonable to assume
that there will be many 1/30 second “moments” when the eye records
substantially higher than the “average” resolution that would be otherwise
calculated on the basis of a model for atmospheric fluctuation.  All that
remains is to postulate that the human brain is capable of integrating
several such discrete instances into a coherent picture of “what is seen”,
and I can convince myself that a visual observer can in fact, outperform the
predictions of the long-exposure model.  

 

Over the years, I have heard many anecdotal reports from experienced
observers claiming to do just that.  Unfortunately, there is usually at
least one piece of critical information that is either suspect or missing
altogether from such reports. Whether the missing piece is a quantitative
description of seeing conditions, mirror figure accuracy (as tested in the
shop), time spent letting the optics come to thermal equilibrium, or even
the experience level of the observer, the point I am trying to make in this
(admittedly) rather long-winded post is that in the presence of theoretical
controversy, it is the data from meaningful experiments that should decide
the question.  

 

Now I realize that anything like a controlled experiment designed to answer
the question at what point (figure accuracy vs. aperture) should a visual
observer stop worrying about optical quality issues is probably not
achievable by a group of individuals scattered across time and space. Hence
I would like to propose an activity that might in fact be achievable; an
on-going verbal sharing of experience guided by a few precepts that (I hope)
we can all agree on.

 

1.	Some attempt should be made to quantify and communicate the
atmospheric seeing limit FOR THE APERTURE BEING USED to report the
observation.
2.	Some attempt should be made to quantitatively understand and
communicate the optical figure accuracy of the instrument as bench tested.
Even better would be to describe the test method (i.e. knife edge,
interferometry), how many data samples were obtained, and what, if anything,
was done to measure astigmatism and separate out the effects of mirror
self-weight deflection in the test stand.
3.	The time spent letting the scope come to or near thermal equilibrium
should be noted, as well as whether or not a fan assist was used.
4.	A description of the observer’s experience level, in regard to what
he/she is actually looking at when they make an observation.
5.	Magnification used.

 

I can start us off by recalling observations I have made through several
scopes for which I have a pretty good data trail, at least as far as
describing optical quality on the bench tests goes.  But it is dinner time
now, and I will only do this if others on this list think it would be of
value.  Thanks for taking the time to read this.

 

Scott Milligan

 

 

_______________________________________________
ATM mailing list http://www.atmlist.net/