About 24 of us turned up, despite the cold and rain and dark. Before starting, Chris Douglass (whose name I keep mis-spelling and who I keep calling Michael Sidonio, for reasons that you can only guess at, and greatly to my personal danger) warned us that next meeting (Wed 5 August), the Kitchen will be out of bounds. I will bring a tea urn, fire stick, or so.
Monte Wilson sent his apologies – Baby Sitting Duty. Geoff Smith also sent his fond regards, reporting that he was having to drink wine in the South of France.
Atlantis
Greg Priestley took us through the Atlantis lift-off, a topic especially dear to us since the Atlantis was off to upgrade Hubble. Greg’s photos were astonishingly clear given that he was photographing from six miles away. They packed a powerful emotional thump.
Introduction to Narrowband
Fred Vanderhaven gave two talks back to back. He began with several particularly magnificent and well-thought-out slides, which summarized the logic behind narrowband imaging. He explained how emission nebulas (eg Eagle, Lagoon), planetary nebulae (Helix, etc), Wolf Rayet objects (eg Thor, Toby), and Supernova Remnants (Veil, Pencil, etc) fluoresce at several very specific wavelengths including those of H-alpha, Oxygen-III, and Sulphur-II. Conversely stars, (and hence galaxies, clusters, etc), incandescent lamps, and moonglow show a continuum spectrum which is almost totally rejected by the narrowband filters. Street lighting is also strongly rejected. Thus EM’s, PN’s, WR’s, and SNR’s can be highly selectively photographed in narrowband, even from the suburbs and in strong moonlight. He warned that normal unmodfied DSLR’s are unable to record S-II, as it is almost infrared, and are poor at recording H-alpha, which is also extremely red. Even O-III tends to fall between the blue and green phosphors. Thus one ideally needs a monochrome camera. Even with a monochrome camera, since most of the light is rejected by the filters, extremely long exposures are required – eg up to 10 hours per filter. However, Fred went on to show that even with a DSLR, narrowband images can be taken (especially in H-alpha and O-III) so long as extremely long exposures are used. The result is images showing extremely fine detail and extremely low background noise, even from the suburbs. Fred further explained how to combine the three monochrome (ie B/W) narrowband images into a single colour image, by randomly assigning colours (eg in PhotoShop). Since H-alpha and S-II are both naturally a very deep red, it makes no sense to try assigning “natural” colours. A “standard” and pleasing mapping is the Hubble Palette, where H-alpha is randomly assigned green, O-III is randomly assigned blue, and S-II is randomly assigned red. Fred closed off this first section by pointing out how if one does use a single-shot colour camera (eg a DSLR), one still needs to make these artificial assignments. The “natural” colours just won’t do.
The second part of his talk was on how to combine a narrowband image (eg 5 hrs of H-alpha of the Lagoon in his example) with a much shorter, more blurry, traditional RGB, to produce a colour image with the same sort of detail as an H-alpha image. He showed two methods. The first was very easy and intuitive. After doing the usual align, stack, and stretch on the H-alpha image, to produce a single H-alpha “answer”, and similarly after the usual processing on the RGB stack, open the H-alpha image in PhotoShop. Now cut and paste the RGB image over it as a layer. Assign the RGB image the layer property “colour”. Hey, Presto! you now have the colour from the RGB image and the detail from the H-alpha image. Fred concluded (under rude and insufferable agitation to finish up from the Chair) with a much more difficult embellishment, in which the fat and bloated stars from the RGB image were selected using a selection mask, and then Deep Magic performed using a Minimum filter. Sadly, even Fearless Leader was unable to follow the details of this section, though there were wise and appreciative nods from the direction of Mick McClough and others. (Fearless Leader normally handles this step with vicious deconvolution of the RGB image).
Fred is a red-hot expert on narrowband imaging, with images on NASA’s POTD. We hope to persuade Fred to give further very short talks on highly specific and precisely timed aspects of this huge topic over the coming months, as he has much to teach us.
Focussing: Theory
I started off showing how a diffraction limited F/4 scope needs to be focussed to within 16 microns of perfect in order to achieve less than ¼ wavelength of distortion. Recall that for a typical Crayford style focusser, one revolution of the fine focus knob is 1250 microns. So to get diffraction limited focus on an F/4 scope, one needs in theory to be within 1/80 of a revolution of the fine focus knob – a half a touch! For an F/8 machine, the figure is 64 microns, or a more manageable 1/20 of a revolution of the knob.
Later, I addressed the question, “How accurate do you really have to be?”. I showed a graph of FWHM as a function of error in focus, and showed how on my F/8 2540 mm focal length PlaneWave and the STL-11000M chip, mis-focussing fattened the FWHM by 4.4 pixels per millimeter. Thus to be within half a pixel of best focus, one needed to be within about 100 microns of best focus – about 1/12 of a turn of the fine focus knob. On an F/4 machine, one would need to be four times more accurate. These “real life” figures are still pretty tight.
I also gave a rather rushed and incomprehensible demo of how one could use a diffraction grated 3-hole Hartmann mask and some software (which members are welcome to have for free) to find best focus within 35 microns on a good night, more than adequate for my machine, by either the theoretical or the practical analysis.
I made the point that when working to that sort of accuracy, seeing had a much larger effect than focussing error. Mark Trudgett correctly pointed out that was not a cause for complacency, and that it all added up: every little bit counted. I agree.
Focussing Workshop
Here we solicited group experience. The results were interesting and surprising in their variation.
About half of us routinely use a Hartmann or Bahtinov mask. The remainder mostly use the “eyeball” method, either with test shots or with a live preview mode. (Michael Samerski recommends a magnifying glass for the live preview on a DSLR.) Fred pushes a button and his machine focusses itself in ten seconds. The rest of us take between a few minutes and up to half an hour to focus. (I take about 10 mins). Nobody much seems to refocus during the night. I strongly suggest that for people with fast scopes (eg F/4 or F/5) that on a long and freezing Wirruna night with excellent seeing J they watch whether their stars sometimes get fatter as the night progresses, in which case they might benefit from refocussing every few hours.
Apart from myself with my Third Harmonic method, nobody uses quantitative methods for focussing (eg FWHM, peak brightness, faint star count, etc), because they take too long. Again, apart from Fred and myself, software assisted or fully automatic focussing was not popular.
When we actually came to look at some images during the second half of the night, some of us wondered whether perhaps this topic of focussing might be reviewed in a little more detail on another occasion. Although mostly the images were of the highest quality, there were also some fat stars about.
Image Feast
It was wonderful to see the number of superb and interesting and varied images produced, despite the ghastly weather over the last 2 months. It was also wonderful to see the large number of people bringing their images along on a memory stick, which meant that we were able to see pictures by MBJ, Marc, Mark, Joe, Matt, and Chris (who holds dual citizenship with Macarthur Soc). All were interesting, most were spectacular, and the presenters worked quickly and efficiently, resulting in something of a recent record in number of images shown and commented on in detail. Chris threw in a few shots of the moon, a neglected object. It is also very nice to see that people are starting to talk a little about the images – what they are showing us, how they were taken, why they chose that object, what it is astrophysically, what interesting features are lurking in the corners. This is good.
Of particular relevance to Fred’s talk was the superb narrowband images taken by Marc, using Very Economical Equipment from a bright suburban sky under the full moon. Although he didn’t say so, Marc has worked extremely hard at improving the mechanical alignment and optical collimation of his scope, and it shows: the field is now far flatter and there is far less coma than ex-factory. He himself pointed out that flocking his scope (with a commercial black velvet-like self-adhesive material) produced a stunning increase in contrast, with perhaps 2/3 of the light previously reaching the chip being stray.
Next Month
I have been told to slow down, pack less stuff into a night, and leave more time for discussion. However, a “respite night” of pure Image Feast was not felt to be required just yet.
I therefore propose to give a fairly short talk on deconvolution, showing how one can produce a spectacular reduction in star size, and pull the finest detail out of nebula and galaxy alike, without producing a gritty background. We can talk a bit about Van Cittert (for the moon) and Richardson-Lucy (for faint fuzzies), about Kernels and where you get them, and Relaxation Curves, and perhaps Cabbages and Kings. I also have a little Fable or Fairy Story involving a Shady Gentleman in a Disreputable Trench Coat, selling Photographs, to illustrate the mathematics and theory.
Please email me as soon as possible if there is a topic you would like to talk about, or if there is a topic that you would like to hear discussed. Also any feedback on how you would like the meetings to be run, or how we could do it better, are solicited.
Hopefully, the winters of yore will return, and we will have some good images to show.
Best,
Mike Berthon-Jones
Leonay Letterbox Observatory
Lost Photon Patrol