On Aug 14, 5:41 pm, Jim Mosher <
jimmos...@gmail.com> wrote:
<snip all>
Jim,
The main take away point is that hyper-exposure to bring out faint
otherwise _visually_ invisible detail provides an alternative means by
which impact imagers may practice before the actual October impact.
It presents an alternative means to the previous consenus practice
method of finding the faintest star that can be imaged as it is
eclipsed by the bright limb or near the poles.
In this regard and with respect to the acknowledgements stated below,
the central warrant of the prior post holds. Imaging of faint limb
objects not otherwise visible with the naked eye illustrates that CCDs
provide an additional photographic magnitude reach beyond the visual
that is not completely overwhelmed by the glare from the bright limb.
That is an unremarkable and expected qualitative note on lunar
imaging.
To complete analysis of this main point, the next question is
quantitative - how faint are faint features recovered through
imaging. The rough preliminary answer is an additional 1.5 mags from
the mid-tones of a lunar image. However, to complete an analysis, a
nearby open cluster would have to imaged at the same setting and the
same approximate altitude as the Moon is imaged. During the Sept.
9-11 lunation and imaging opportunity, M45 and the Moon will be
favorably positioned to gather CCD photometry data to answer such a
question.
> The attachment may be helpful for correctly identifying the
> features you see in what you call the "Malapert shadow hole"
Your criticism that I have improperly identified the faint background
features is a correct and is acknowledged. I suspect the faint
features in the background of the Malapert dark hole are part of
either a rim of Shoemaker or the highlands between Shoemaker and
Shackleton. See site line markups on supplmental panel image:
http://members.csolutions.net/fisherka/astronote/observed/LCROSS/20090809_1123UT/20080809_1123UT_OPW111d52mN40d46mKAF9B_DiscussionPanel.jpg
http://tinyurl.com/mlmseo
> The reason your
> camera could detect them only with increased exposure would seem to me
> to be a matter of lower resolution (smearing them out into darkness)
> coupled with some difference in the linearity settings/processing
> controlling how low light levels are displayed. It seems unlikely to
> me that the success of prolonged exposures in bringing out additional
> features demonstrates indirect illumination of those features by
> sunlight scattered from nearby peaks and ridges: all the features
> brought out by the deeper exposure would have already been present at
> the lower sun angle photographed by the AstronoMinsk images. And
> there is no reason to think the AstronoMinsk image is showing anything
> other than direct lighting.
Your criticism is correct that the faint features in the background of
the Malapert dark hole are in direct sunlight and are not illuminated
by backlighting as I originally supposed.
However, the main warrant of the exercise and purpose of the analysis
still holds. The context of this test was to find alternative to the
practice imaging of faint objects near the poles to the method of
finding the faintest star that can be imaged as it is eclipsed by the
bright limb or near the poles.
The faint features in the background (with one exception) were not
detected _visually_ at 475 power (1900 fl / 4mm ep). They were
however brought out by CCD imaging. And they were imaged even though
the area is framed by fully illuminated lunar surface on the bright
side of the terminator. Glare is not preventing the capture of the
additional detail.
In the attached discussion panel, I have included a raw jpg export on
one fits frame and labeled the frame with the raw ADUs for six
individual pixels.
The middle frame is a Goldstone radar based topographic map oriented
using your LTVT sofware to the circumstances of Aug. 11. Using, LTVT,
I also gathered sun angle ephemeris data for each of the six pixels.
That data is listed in tabular form as follows:
Pixel region - ADUs - x-y coord in raw FITS - emphermis sunangle at
pixel
Leibnitz beta bright spot - 27692 - x180y206 - SA 3.6°
Shoemaker back rim moderate spot - 7690 - x234y184 - SA 0.3°
Shoemaker back rim faintest spot - 6243 - x218y186 - SA 0.3°
Moretus central peak bright face - 31564 x266y426 - SA 15.7
Dark half of Malapert E - 13242 x156y206- SA 3.2°
Simplieus J - 32466 - x96,y260 - SA 8.9°
While this is a clear filter image without two filter transforms, raw
differential magnitudes based on the ADU counts indicates that
relative brightness (27692/6243=4.4) between the bright spot on
Leibnitz beta and the faint line in the background (at pixel x218y186)
is about 1.6 magnitudes.
Similar photometry cannot be extracted from the technically superior
Astronominsk and Groell images because those are probably post-
processing, gamma enhanced and stacked images. Similarly, we have no
information from the raw source images or whether when those images
were taken, what features where or where not _visually_ as opposed to
photographically detected.
You may be making an interpretative error by not accounting for the
difference by visual observation and photographic seeing. That faint
detail is captured on gamma enchanced images by Astronominsk and
Groell does not imply that all of the faint detail was or could have
been observed visually.
The above magnitude computation is a relative measurement. To really
be useful for ascertaining whether the LCROSS ejecta curtain can be
captured by imaging, an rough absolute magnitude of the features at
various sun angles would need to be determined. A suggested procedure
for that will be discussed in a follow-up post.
Nonetheless, the technique illustrated is both relevant and highly
valuable as a pre-event practice method for amateurs desiring to image
the LCROSS image. The faint features in the Malapert dark hole are
illuminated by extremely low elevation sunlight - 0.3 degs in my image
and 0.2 degs in the AstronoMinsk image. That they are faintly
illuminated by very low level sun angle light verses backlighting is
not fatal to the practice technique. The point is they are much
fainter than the foreground and therefore are useful for imaging
practice.
The LTVT software engine can be useful in identifying craters at
either pole on almost any night of a lunation with analogous
illumination at extremely low sun angles. Practice imaging becomes
seeing how far into or near the terminator that the faintest feature
can be recovered out of the _visual_ blackness.
This practice method is less time constrained than trying to find the
rare star of the right magnitude (between mag 7 to 0) to graze or slip
past a pole and then taking an image as it passes.
> [W]hy the main disk of the Moon turns
> black when you increase the exposure, giving a "solarized"
> appearance. Is some kind of digital counter overflowing? If so, why
> is it stuck on black and not showing another cycle of gray levels?
The effect is camera and software related and may be specific to the
Meade DSI Pro One used. The bottom frame of discussion panel includes
the final stacked image with a histogram from one frame of the stacked
sources fits images. When the Meade DSI Pro exposes past its well
capacity, it offsets the pixel value by minus 2^16 (65536) - making
the pixel value negative. Since zero is the value of a black pixel,
overexposed pixels and those stacked pixels that exceed 65536 are
rendered black by Registax. In the final image, negative pixel values
are removed by clipping.
> "Age" and "percent illumination" (of the Moon as a
> whole) may be helpful for orientation, but to the extent
> they dictate looking at some other date they are irrelevant.
Your conclusion does not follow from the observing context.
The remaining analogous date (9-10) suggested by Rick Bladridge is not
a precise replication of the conditions at the time of impact. The
libration in latitude is significantly different (-5.4 on 9-10 verses
-3.5 on 10-9) than what will occur at the time of impact. Similarly,
the libration in latitude condition on 8-11 significantly different
(-5.9 verses -3.5 on impact day).
Rough analogous dates are not irrelevant due to this lack of
precision. Imagers, particularly DSO amateurs with little high-
resolution lunar imaging, who will be successful are those that
practice imaging faint features surrounded by the much brighter
terrain on the limb. Location and tracking of the correct lunar
feature at f/30 and higher is not a trivial task. Practicing good
exposure of faint features should occur without regard to
pragmatically unattainable precision of libration in longitude and
latitude that will exactly replicate the conditions of impact. On the
date of the impact, the faintly illuminated background features will
not be visible but fortunately will be visible one last time around
Sept. 10 should imagers wish to practice their exposure technique.
Again, Jim's LTVT software engine can be useful in identifying craters
at either pole on almost any night of a lunation with analogous
illumination at extremely low sun angles.
As can be seen by comparing the top and bottom images of the
discussion panel, hyperexposure with stacking can bring out
considerable more relative faint detail and more detail than will be
seen _visually_. That the Gorell and Astronominsk images used larger
apertures and better equipment misses the central point. The
important speculative question is what relatively more faint detail
could they have brought out with their superior equipment had the
objective of their imaging sessions been to purposefully try to more
faint image detail at the south pole.
> Also, your raw image "20080809_1123UT_OPW111d52mN40d46mKAF0_SP.jpg"
> (shouldn't it say 20090809 ?) appears to be compressed vertically
> compared to the geometry one might have expected on 9 Aug 2009. Is it
> using non-square pixels?
Your criticism that the image has a processing error is correct. Meade
DSI Pro I wells have a rectangular shape. I believe the error is that
I rotated the image before I corrected for the rectangular pixels. It
should be done the other way around - correct for the pixel size first
and then rotate.
> The general public should have little trouble finding the south pole simply by telling them to look for where terminator crosses the "battered" end.
This reasoning about the abilities of the general public errs based on
experience. The most likely scenario will be that the inexperienced
general public will start observing two to four days before the impact
date to get oriented to the lunar disk and in response to pre-impact
publicity. At that time, the terminator's "battered" end does not
disappear at the south pole and illumination creates a false south
pole dark hole asterism that looks very similar to the Malapert dark
hole, but the east of the true south pole along the limb. One lunar
observer here was fooled by it and made a "south pole" image east of
Scott along the lunar limb. When I began imaging on August 7 - being
hurried after a long-day at work - I made exactly the same error and
made a wonderful image of the wrong spot on the limb - and I'm no
lunar neophyte. that was the motivation in devising the supplemental
site light technique. Closer to the full Moon, the Moretus, Simpelius
and Schomberger landmarks are almost invisible in the brighter flat
lighting.
A better star party mentoring technique would be to use site lines
that I suggest, because they can be traced to the correct spot on the
lunar limb even at high levels of illumination. The traditional route
that you suggest (as has been discussed in this newsgroup for
sometime) is appropriate for star party mentoring on the evening
before and the night of impact. But at higher levels of lunar
brightness, the Morteus siteing route does not work very well.
On August 14, the Moon traveled through the apparent center of the
Pleiades cluster. Due to a cloud-out and technical problems, I was
unsuccessful in capturing a south pole image and a magnitude
calibration image of Alcyone. Nonethless, the exercise suggests a
simpmle technique to establish rough absolute magnitudes for the
bright and faintest features recoverable at the south pole with a CCD.
I will discuss the August 14 test run in a follow-up post.
Clear Skies - Kurt