Michael A. Covington    Michael A. Covington, Ph.D.
Books by Michael Covington
Consulting Services
Previous months
About this notebook
Search site or Web

Daily Notebook

Popular topics on this page:
Clearer thinking about character sets and Unicode
Unicode or not Unicode?
Why older adults are vulnerable on the Internet (II)
Putting your own horizon line into Stellarium
"Feature transfer error" when installing software
Some administrator accounts can't install software
Notes about flat fields in astrophotography
How much does it cost to get a book published?
Using Stellarium for locations other than your own
Canon 40D and 60Da spectral response compared
Note on making a red LED penlight for astronomy
Photoshop "not enough memory" opening DeepSkyStacker file
Dark nebulae near Altair
Dark nebulae near Delta Aquilae
Dark and bright nebulae near Mu Cephei
Dark nebulae in Scutum
North America Nebula (NGC 7000)
Nebulae and star clouds in Sagittarius
Nebulae near Gamma Cygni
Star clouds and dark nebulae around M7
Andromeda Galaxy (M31)
Veil Nebula
Many more...

For more topics, scroll down, press Ctrl-F to search the page, or check previous months.

For the latest edition of this page at any time, use this link: www.covingtoninnovations.com/michael/blog

Note: The ads that you see here are not controlled, sponsored or endorsed by Covington Innovations.
They are brought to you by Google and will vary depending on your personal browsing history.



Yes, we have sunspots... this particular solar maximum is proceeding irregularly, but there are plenty at the moment. This is a stack of four 1/2000-second exposures with my Celestron 5 and Canon 60Da, processed with RegiStax. I used a Baader solar filter that has seen better days — there is actually a hole in it, patched with tape — and have just ordered a new Thousand Oaks filter. (I had one back in May 2012 but it was the wrong size for my telescope.)

Incidentally, I am such a coward about solar observing that, these days, even though I know the filter is perfectly safe, I never put my eye to the eyepiece, relying entirely on the camera's video screen instead.


Clearer thinking about character sets and Unicode

Here are some further thoughts about the Unicode question. I've been trying to plot strategy for several natural language processing projects, and I finally realized that before choosing a file encoding, I needed to make some decisions about character sets.

Unlike the file encodings, the character sets form a neat hierarchy, each one subsuming the previous one. Specifically:

  • ASCII has the minimum for encoding typewritten English.
  • ISO Latin-1 (ISO-8859-1) adds the accented letters needed for major western languages. Each accented letter has a unique code.
  • Windows-1252 adds curved quotation marks, dashes, and a few other specialized characters. Reportedly, most implementations of Latin-1 nowadays are actually Windows-1252.
  • 16-bit Unicode (that is, Unicode without code points above FFFF) adds a vast array of characters for all the languages of the world, including many non-western writing systems. Each character fits in a 16-bit char type in modern programming languages. However, some characters can be written more than one way (the combining diacritic problem) and some characters simply occur more than once due to overlapping character sets that were combined to make Unicode.
  • Full Unicode includes some really exotic characters that do not fit in a single 16-bit char.

Now here are my thoughts:

Natural language processing software needs to control its character set so that recognition of words will be easy and reliable.

The input to our software should be able to use any popular encoding, with UTF-8 as probably the preferred one.

Upon input, as a minimum, Unicode should be normalized so that each character is always written the same way.

In fact my inclination is to impose heavier limitations, even possibly reducing all characters to ASCII after tokenization. (Accents are often left out of headlines, and relying on them would lead to mistakes. Dashes can be tokenized as multiple hyphens, as on a typewriter. My inclination is not to use non-Latin alphabets or writing systems internally, to avoid putting an obstacle in the way of colleagues who don't know an exotic language. That is, if a non-Latin alphabet is involved, the tokenizer should transliterate.)

The output of the software might well be pure ASCII, or, for legibility, Windows-1252 or UTF-8, user-selectable. It should be confined to the Windows-1252 character set at the most.


Unicode or not Unicode?

[Extensively revised.]

I've been matching wits with Unicode the past couple of days, getting ready for some projects that involve foreign languages, and have been a bit vexed by it all.

ASCII is a thing of the past. Even the English language requires non-ASCII characters, such as curved opening and closing quotation marks. Languages like Spanish certainly require non-ASCII characters.

ISO Latin-1 (ISO-8859-1), which is essentially the same as Windows "ANSI" (Windows-1252), has some advantages. Every character is one byte. The characters commonly needed for English, Spanish, French, and German are all there. Every character is written only one way (we do not have the Unicode "combining diacritic" problem).

But ISO Latin-1 doesn't have all the characters of all the languages of the world.

There are several ways of encoding Unicode, but UTF-8 is the only serious contender, in my view. So it boils down to Latin-1 versus UTF-8.

Until just now, I was gung-ho to use UTF-8 for everything, and to include the byte order mark on every UTF-8 file.

Then I started looking at how various pieces of software respond to various kinds of text files. Major pieces of Windows software, including Notepad and Excel, prefers Windows-1252 but will automatically accept UTF-8 if there is a byte order mark.

R, on the other hand, takes either Latin-1, or UTF-8 without a byte order mark. This is the case under both Windows and Linux.

In fact, Latin-1 a.k.a. Windows-1252 is better supported in Linux and on the Macintosh than I thought. It seems to be the character set that won't die...

So my considered opinion, now, is that UTF-8 is best but Latin-1 is tolerable. I also think we should consider using the extension .utf8 instead of .txt for files that are known to be Unicode.

Whether to include the byte order mark, I don't know. If it's there, many pieces of software (including the input libraries of C#) will automatically recognize the file as UTF-8 and will automatically skip the byte order mark. But R reads it and keeps the byte order mark as if it were a character. Can't win.


Why older adults are especially vulnerable on the Internet (II)

Back in July I wrote about why mature adults are especially vulnerable to scams, viruses, malware, and blunders on the Internet. If you haven't read it, click through and do so. Then come back to here...

I've noticed a few more factors at work.

One of them is that older adults often underestimate the size of the Internet and think only a few like-minded people will ever communicate with them, even indirectly. Sometimes they even think anything they receive must come from a close associate. Not at all! The Internet puts you in touch with the wide world, and it's very wide indeed.

I call this the "small-circle-of-friends illusion" and it affects newcomers to the Internet of all ages, but I've seen older people stick with it when they should have known better.

This illusion often makes mature adults far too willing to trust, and pass along, a message that came from nobody-knows-where, nobody-knows-when. Maybe someone lost a puppy and asked all the friends of their friends to spread the word. But for all you know, the puppy was lost 3 years ago, 2000 miles away, and was found a week later. Just because it came to you from your best friend doesn't mean it originated in your small circle!

Bottom line: "Pass this on" is not the right way to spread information on the Internet. It's like mailing someone a letter and asking him to make copies by hand and mail them to all of his friends, when what you really needed was to put something in the newspaper. The Internet's equivalent of the newspaper is the World Wide Web, and anybody can write on it. "Pass it on" e-mails are not the right way to spread news. Yes, you feel like you're helping, but the feeling is an illusion; either you're doing no good, or you're being manipulated.

A second vulnerability is that some, not all, older adults are unwilling to admit that they don't know enough. They take a guess about something important, and are satisfied with it, and soon forget that it's only a guess.

These are typically people who haven't been in a new environment for a long time and are simply not open to learning. If you're not open to learning, in the world of personal computing you're sunk!

The rest of us often laugh at people who won't admit to others (or even to themselves) that they've taken a wild guess. Think of the people who mistake DVD drives for coffee cup holders. You might guess it's a coffee cup holder, but do you know? Of course not. You have to remember the distinction between what you've guessed and what you know from reliable sources.

Even worse things can happen. By now, one of the clichés of the Internet is the older adult (typically a parent or grandparent) who copies younger people's fashionable language (or so they think) without knowing what it means. All of a sudden Grandma is using abbreviations for dirty words, and she doesn't know it!


Extreme fixed-tripod astrophotography: Cygnus

This is a hasty experiment to see if the Milky Way can be photographed with only the equipment that a serious non-astronomical photographer would have. I'm trying to come up with a technique that a friend will be able to use on a trip to Africa next spring.

This is the middle of Cygnus, photographed with a Canon 60Da on a fixed tripod and a 50-mm f/1.8 lens at f/2.8. This is a stack of sixteen 3.2-second exposures at ISO 6400, combined with DeepSkyStacker. Long exposure noise reduction was done in the camera, and no flat-fielding was done (although it certainly could be).

As you can see, it was a success. You can even see a good bit of nebulosity, although a camera with conventional spectral response (rather than a 60Da) might not pick it up.


Photoshop reports "not enough memory" opening DeepSkyStacker output

"Could not complete your request because there is not enough memory (RAM)."

The autosave.tif file produced by DeepSkyStacker is in 32-bit rational-number TIFF format. Photoshop (version 5.1, suite 5.5) can open this if you're using the 64-bit version of Photoshop, but not the 32-bit version. (Adobe supplies them together, but the 64-bit version of Photoshop installs and runs only if you are using a 64-bit version of Windows.) If you try to open this type of file with 32-bit Photoshop, you get the misleading error message shown.

(I have no idea whether 32-bit Photoshop does not support that type of TIFF file or whether there is perhaps something subtly wrong with the format of the file itself as created by DeepSkyStacker. I have heard reports of other software that also will not open these files. It is significant that the Photoshop error message is misleading, suggesting that the file is malformed in some subtle way. If Photoshop failed to recognize the file as a supported format, it should say so.)

One workaround is to save your file from DeepSkyStacker as 16-bit TIFF, preferably after doing some adjustment of levels and curves.

My experience is that PixInsight has no trouble opening these files. I am using the 64-bit version and have not tested the 32-bit version.

I want to thank Brian Johnson for asking this question on the DeepSkyStacker mailing list, and Dan Crowson for pointing out the answer.

The government is running again, I have a new consulting project, and I'm about to stop having quite so much time for astronomy. Stay tuned for a much wider range of topics in the Daily Notebook, and perhaps a bit less material per day.


Note on making a red LED penlight for astronomy

I've made myself a red LED penlight to use while doing astronomy, red so that it doesn't spoil my night vision. I don't have time to write up a complete "how-to," nor do I consider it quite perfected, but here's how I proceeded.

(1) Get a penlight of the kind doctors use, such as this one. Important attributes: Replaceable batteries and bulb; old-style penlight bulb (not LED); bulb is somewhat recessed so the bright light is not visible from the side (important for astronomy).

(2) Also needed: A bulb with the same kind of threads on its base, but preferably with a round globe; a red LED; a 100-ohm resistor.

(3) Soak the bulb in acetone or maybe lighter fluid to dissolve the cement that holds the globe to the brass base. Pull the globe free, unsolder it from the base (after the flammable solvent has evaporated!), and clean out the base.

(4) Carefully reassemble the LED and resistor into the bulb base. After soldering, I used plenty of glue and also ended up using part of a small rubber grommet to keep the LED centered and to give it a "shoulder" so it wouldn't protrude too far.

(5) Reassemble the penlight and try it out.

I've done 2 variations on this project over the years, and neither one came out quite perfect enough to publish, although both are serviceable. I may work it up into a construction project article later.


Canon 40D and 60Da spectral response compared

The Canon 60Da is a special astronomy DSLR with extended sensitivity to the red part of the spectrum (especially 656 nm, the wavelength at which most hydrogen nebulae shine). The 40D is a conventional DSLR. Here is a comparison of what they see, looking into a Project STAR spectrometer.

The light source is a white piece of paper illuminated by a halogen incandescent lamp; there were also fluorescent lights on in the room, and you can see some of their bright emission lines, as confirmation that the wavelength scale is correct.

You can see that the 60Da has useful response out to 690 nm; the 40D, like all normal digital cameras, stops around 640 nm to better simulate human color vision.

Important: The blue and greenish light to the right of 450 nm is an instrumental artifact from another order of the diffraction spectrum.

I took these pictures quickly so I could include Ektachrome film in the comparison. The Ektachrome hasn't been developed yet, of course, and if I didn't get a successful picture, it will be too late, since it's my very last roll of Ektachrome. But I hope to publish a version of this picture that includes Ektachrome before too long.


Veil Nebula (Cirrus Nebula)

Here's the last of my pictures from Deerlick on October 5. Same technique as all the recent ones, but this is a stack of 17 (yes, 17) 30-second exposures. You're looking at a rich field in Cygnus with the Veil Nebula (a supernova remnant) at the lower left — a C-shaped portion is the most conspicuous part. The star cluster NGC 6940 is at the lower right, and there is a patch of dark nebula above the center.

Now I must either take more pictures or find something else to blog about!


Andromeda Galaxy

This isn't my best-ever picture of M31, but it's notable for the seemingly inadequate equipment with which it was taken: a 105-mm f/2.8 lens and Canon 60Da camera body on an iOptron SkyTracker. This is a stack of eleven 30-second exposures at ISO 3200. You're looking at an enlarged area from the middle of the picture.


Nebulae and star clouds around M7

This rich star field, with star clouds and dark dust clouds, includes the star cluster M7 (Ptolemy's Cluster, above center) and even some emission nebulae, including the Cat's Paw Nebula at lower right. Same technique and equipment as the previous picture (below), but only 8 exposures were stacked.


Bright and dark nebulae around Gamma Cygni

This rich star field, with red hydrogen nebulae and dark dust clouds in front of clouds of distant stars, is in the middle of the constellation Cygnus. I photographed it from Deerlick on October 5. Stack of fifteen 30-second exposures at ISO 3200 with a Canon 60Da and Sigma 105/2.8 lens, wide open, on an iOptron Skytracker.


Using Stellarium for locations other than your own

Today I was trying to use Stellarium to compute the appearance of the sky in Zambia, to help a friend prepare for a trip. Everything looked right except that the times were a few hours off — the sun seemed to be setting at noon, or something like that.

I figured out what was going on: something that was logical but produced surprising results. Stellarium asks your operating system (Windows, MacOS, or Linux) for the true time of day and your time zone. So it was correctly showing me the sun setting in Zambia at noon Eastern Daylight Time, which was, I think, 6 p.m. there.

There are several workarounds, but the simplest is to set Windows itself to your intended location when you want to use Stellarium to predict the sky somewhere else. Then the computer's clock will show the time zone of the place you're trying to calculate.

I said "everything looked right." Yes, I did calculate the Zambian sky roughly in my head before asking the computer. I do things like that... I learned about the sky back when we didn't have computers to do our figuring for us. The knowledge has served me well. Practically every time I've looked at the night sky since 1970 or so, I've known what I was looking at. I was only flummoxed once, during a transpacific flight in 1973, when I had been asleep and woke up not knowing whether I was in the Northern or Southern Hemisphere and also not knowing whether I was in the Eastern or Western Hemisphere. I then failed to recognize Fomalhaut.


Martin Mobberley's entertaining, informal biography of Sir Patrick Moore has been an enjoyable trip down memory lane. One of my greatest regrets is that I never met Moore; I've only been to one B.A.A. meeting (maybe two?) and he wasn't there. But the height of his career coincided with my entry into amateur astronomy, and also Martin Mobberley's, so the book called forth a host of memories.

I am about to try out PixInsight image processing software. It comes from Spain, and one thing that encourages me is that I see plenty of excellent English and also excellent Spanish on its web site. Far too many bilingual endeavors end up mangling one of the two languages, or at least peppering it with errors. These people looks like perfectionists — exactly what we want.

OK, government...

We will not soon forget the damage done by our Congress, which has just now agreed to pay its bills and reopen the federal government. Even the threat or needless risk of defaulting on federal debts damages our economy and makes us look silly. Well-governed countries do not do such things.

Everything I said two weeks ago still stands. I plan to vote against every incumbent in the next election for their offices, whether 2014 or 2016. (This can be done at the primary level, to avoid changing parties.) There is plenty of blame to go around; both sides were unreasonable; both sides are guilty of legislative malpractice.


Dark nebulae near Delta Aquilae

I made a second trip to Deerlick on October 5, and conditions were not quite as good as October 1. Here's one of the pictures from the second trip, a stack of nine 30-second exposures with a 105-mm lens at f/2.8, Canon 60Da camera body, and iOptron SkyTracker, at Deerlick, processed in the same way as pictures posted earlier this month. You're looking at the middle of the constellation Aquila.

This flurry of astronomical activity is brought to you by the federal government shutdown. As soon as Congress resumes doing its job, I'll be busier!



Putting your own horizon line into Stellarium

I am wholeheartedly recommending Stellarium sky charting software. This is a brief note on how to put your own measured horizon line into it.

The Stellarium people give you plenty of instructions (here, here, and here) about how to create your own picture of the earthly landscape to display under the stars in Stellarium. You can even do this by taking a panoramic photograph, then transforming it into a 1024×2048-pixel PNG file with the sky marked as transparent.

What I want to address is something much simpler. You may have measured the altitude and azimuth of points along the obstructed horizon around your telescope site, as I have. A file containing these data in numerical form is useful with TheSky and Skychart, among others. How can we use it with Stellarium?

Here's what I did. First I plotted the data as a line graph using R, taking care to use the same scale for the horizontal and vertical axes and to have 0 (altitude) in the middle of the graph:

I converted this to a high-resolution bitmap, cropped it at the left and right ends of the zigzag line, and, without changing the shape, converted it into a 1024×2048-pixel PNG file in which the line was yellow and everything else was transparent. Here's how it looked in Photoshop:

Because this is a short note, I can't go into the details of how to do this, but there are plenty of tutorials on how to create transparent areas in PNGs. Essentially you select just the line using the magic wand tool, paste it onto a transparent layer, and then delete any layers that are not transparent.

Finally I created a landscape.ini file per Stellarium's instructions, making sure to include the lines

type = spherical
angle_rotatez = -90

because for some reason Stellarium expects panoramas to start in the east rather than the north.

Here's the result:

The "ground" is transparent and is in fact just a jagged line through the sky showing the approximate limits of my vision past trees and houses.


"Feature transfer error" when installing software  
Some Administrator accounts can't install software  
Error: -2 The system cannot find the file specified

I've finally exorcised a ghost that was haunting my desktop Windows 7 system. The symptom was this strange-looking error whenever I tried to install certain pieces of software — notably MaxIm DL and Cisco AnyConnect. The workaround was to log on a different Administrator account, and then the installation went just fine. Only one account was haunted, and I was almost ready to delete the account.

This afternoon the error plagued me again as I tried to update MaxIm DL, and I tracked it down and fixed it.

It has to do with permissions on C:\Users\username\AppData\Local\Downloaded Installations. Normally, System, Administrators, and the individual user all have full control. In my "haunted" account, though, only the user (owner) had full control; the other permissions weren't inherited. I think this may have happened when I copied some AppData folders from an older version of the account.

This matters because InstallShield runs as the individual user but the .msi installer runs as System. Thus, the files unpack successfully and then InstallShield hands them off to MSIEXEC, which can't see them. Nor does MSIEXEC tell you what it's looking for. I had to use Process Explorer to find out.

To fix the problem, right-click on C:\Users\username\AppData\Local\Downloaded Installations, and choose Properties, Security, Advanced, Change Permissions. Remove the explicit permission for the individual user, and check "Include inheritable permissions." Then click "Apply," make sure three permissions are showing up, and click "OK."

I found several other folders in AppData with the same peculiarity and fixed them the same way, just in case it matters.


The North America Nebula

North America Nebula

Musicians have theme songs — favorite songs that they sing or play as a sort of personal trademark — and I have the North America Nebula. I never tire of photographing it with different kinds of equipment under different conditions. Here's what I got on October 1, a stack of 10 30-second exposures with a 105-mm lens at f/2.8, Canon 60Da camera body, and iOptron SkyTracker, at Deerlick, processed in the same way as pictures posted earlier this month.

The richness of our galaxy


Here is a very rich field of stars, clusters, and nebulae in roughly (not exactly) the direction of the center of our galaxy. Equipment and technique exactly the same as for the North America Nebula (above). Sagittarius was low, close to the horizon, and you may be able to see some man-made skyglow toward the bottom of the picture. Objects of interest here include the Lagoon and Trifid Nebulae at the lower right, the Swan Nebula at the upper left, and the star cloud M24 above center.



OK, all the astrophotos on this page (for October) have been reprocessed with better flat-field correction, and you should scroll down to see them. I'll start releasing more photographs in the coming days, and I'm also going to reprocess some pictures on the September page. [This has been done.]

The problem was that "set black level to 0" in DeepSkyStacker was checked and should have been unchecked. As a result, the flat-field correction was not quite right; there was a difference in color and contrast between the center of the picture and the edges. Also, the flat-field images that I was using for correction were underexposed.

For the uninitiated: Almost all camera lenses produce a picture that is slightly darker at the edges than at the center. This is not necessarily a defect; it is also a consequence of geometry (the cosine-4 law). But in astronomy, we increase the contrast so much that this needs correcting. The tactic is to take pictures of a blank white surface using the same lens, then perform some computation. The computation was being done incorrectly, and now I've corrected it.

Scroll down, and enjoy!


How much does it cost to get a book published?

I just heard a radio commentator advising a young woman who was thinking of borrowing lots of money to pay to have her first book published. Fortunately, he steered her clear of it. But it got me thinking about how little most people know about the publishing industry.

As a book author, I fairly often get asked two questions that strike me as strange. One is, "How much did it cost you to get your book published?" The other is, "How many millions have you made from your books?"

The answer to the first question is that I have never paid a cent to a book publisher. Never have, never will. The way it works is that the publisher takes all the financial risks, bears all the production expenses, edits, typesets, prints and sells the books, and pays me royalties.

(That, by the way, is why you can't buy my books from me. I don't manufacture or sell them myself.)

The answer to the second is that the income from my books is measured in thousands (or even hundreds), not millions. Not every book is a best-seller. Tens of thousands of good books are published every year, and only a handful make millions. Writing is not the way to get rich.

And that's a problem with some starstruck aspiring writers. Even if you're very, very good, the odds are that you will never be one of the top ten people in the world in your field. The world is big, and ten is a small number.

I would go further and say that the words "I want to be a writer" are a bad sign. "I want to write about X" is much, much better.

Getting back to how books are published: Yes, I know about self-publishing. In fact, it is easier than ever to publish your own book through Amazon, Barnes and Noble, or other companies. They have an efficient distribution network, which is what self-publishing "vanity presses" lacked in the past.

Self-publication doesn't cost much, provided you can provide a well-edited book (perfectly spelled and punctuated, with page numbers, neat layout that follows standard requirements). If you can't do that, you'll have to pay someone to edit the book for you — and in fact I recommend bringing in an editor, because they will catch things you didn't, but you shouldn't give them garbage to work with. If you cannot word your sentences clearly, spell correctly, and punctuate correctly, then in my opinion you are not a writer. You need a writer. You are like an "artist" who wants someone else to touch up the paintings.

Self-publishing is a great way to go if (1) you can afford the (small) cost, and (2) you have something that you know will sell 50 copies, and you want to make it available to the wider public, and (3) you have ways of publicizing it by word of mouth (especially through online friends and web sites that are interested in the subject — but please don't become a spammer).

Don't expect to make money that way. Right now I'm looking at a self-published book (not my own) on Amazon that has apparently not sold any copies at all for several weeks. It's fiction, of course. And here's a related point: Most fiction writers are not as good as they think they are. The fact that you enjoy it doesn't mean anyone else will enjoy it. If you write non-fiction, as I do, it's easier to get feedback from people who will tell you whether the book is interesting and useful. Fiction is a wild stab into the unknown. Even some of the best authors have been rejected by some publishers and critics.

Of course, you can make money self-publishing. If you have a highly salable book, you can make more money than by going through an existing publisher. Why? Because you're doing two jobs and running two businesses, author and publisher. Being a publisher can be a lot of work! It's a business I chose not to go into. I have plenty to do already.

The case against self-publishing is that you don't get an established publisher's prestige. When my books are published by Cambridge University Press, the public has some assurance that they've been reviewed and accepted by competent people. If I published them myself, you'd have no assurance that I'm not a crackpot.

Thanks to new technology, the cost of printing books is plummeting, and that cuts both ways. Not only does it make self-publishing easier, it also opens up new opportunities for commercial publishers. The other day a major scientific publisher came by and basically wanted to publish any research I want to share, including working papers (reports of work in progress). This would be a deal that pays me no royalties but makes the work available quickly to the entire world, each copy printed on demand. It's much better than just putting my research on my own web page, which might vanish if I moved away from UGA.

I can think of several scientists who would like to publish some of their work that way, although none of mine quite fits. (The publisher, by the way, was not Cambridge. Think Germany, and mathematics, and horse-heads...)

The bottom line? "If you're a writer, why aren't you rich?" Because almost no writers are rich. Like most work, writing pays just well enough to get people to do it. In fact, at the moment I'm not writing books because I can earn more money doing other things. That, of course, can change from year to year as opportunities come along.

Those reprocessed astrophotos are coming. Stand by!


Stand by for reprocessing

It has come to my attention that the flat-field corrections applied to my last several astrophotos have not been entirely correct. I'm going to reprocess all the photographs I've published for the past couple of months before releasing more. Stand by and be patient...

Astronomy software

I've long been a user of both TheSky and Starry Night astronomy software, both of which are commercial packages. Recently I started looking for comparable software that might be available free, so I could install it without restriction on multiple computers.

I've tried out Skychart (Cartes du Ciel), which looks very good. Interestingly, Skychart is built with the Lazarus compiler, an open-source multi-platform compiler that keeps alive the Borland Delphi language, the first Windows programming platform I ever really liked.

I also hear good things about Stellarium, which I haven't tried yet. Stay tuned...

Addendum: Both are good, and the price is right — both are free! Stellarium has a more sophisticated user interface, but both do their jobs reliably. I am putting both of them on all my computers. The configuration is easy to port from one computer to another; it's in Application Data\Local\Skychart and Application Data\Roaming\Stellarium respectively.


Dark nebulae in Scutum


[Reprocessed 2013 October 12.]

Here you see a rich region of star clouds and dark nebulae in Scutum, with the cluster M11 at the bottom (looking like a fuzzy bright star). Last month, I photographed this same field from in town. This is a stack of nine 30-second exposures at ISO 3200 with a Canon 60Da, with a Sigma 105-mm f/2.8 lens wide open on an iOptron SkyTracker, calibrated against dark frames, flats, and flat darks using DeepSkyStacker.

Some notes about flat fields in astrophotography


(This information will later be moved into my astrophotography pages. I'm starting out with a review of the basics, but there's some new information a few paragraphs down.)

Besides contrast adjustment and dark frame subtraction, I usually also correct my astrophotos with flat fields to make up for unevenness in the light path. This is how you compensate for darkening at the edges of the field (vignetting and the cos4 law of optics) and even dust on the sensor (which would otherwise show up as blotches).

Flat-fielding is optional. Here's what the foregoing picture of Scutum would look like without flat-fielding. You can see that the illumination falls off at the edges. The contrast stretching that we perform in astrophotography tends to exaggerate what would otherwise be an almost unnoticeable effect.


This is not a bad picture at all, but there's a lot more edge darkening than in the corrected version at the beginning of today's entry. And note that this lens does not have noticeable edge darkening in ordinary photographs. It's a full-frame lens on a compact DSLR. Many lenses would perform far worse.

Now for some points that aren't obvious.

(1) Flat fields are images of a featureless white surface. I use a small battery-powered light box designed for viewing 35-mm slides, held right in front of the lens, almost in contact with it. To reduce the light level I put a couple of layers of white handkerchief in between them, stretched tight across the lens.

(2) Flat fields must use exactly the same optics as the astrophotos, including aperture and focus settings, and preferably be taken at the same session. The reason is that everything affecting how much light falls on the sensor — even dust — must be the same.

(3) Flat fields do not have to match the ISO setting or exposure time of the astrophotos. Generally, they can't. I use a low ISO setting (usually 200, the lowest my camera will permit) and use the exposure meter in the camera to choose an exposure time, exposing about 3 stops more than the meter says. You want a full frame that is appreciably above mid-gray but does not hit maximum white.

(4) Along with flat fields, you need flat darks. These are just like flat fields but taken with the lens cap on; same camera settings, same exposure time. The reason here is that your camera has bias — pixel values do not start at 0 for black, but at some higher number — and your image processing software can't interpret the flat fields correctly unless it knows the bias.

(5) With DeepSkyStacker, all the flat fields and flat darks have to have the same exposure time. Otherwise they won't be recognized as a matching set (and probably won't be valid for doing computations anyway).

(6) The color of the flat fields doesn't matter as long as it isn't strong. A bluish or reddish tinge is OK. You can take flats on the twilight sky using a handkerchief in front of the lens or telescope.

(7) Flat fields should be exposed generously, about three stops above a midtone as determined by the exposure meter.

(8) Flat fields should be taken at slow shutter speeds. At 1/8000 second, my DSLR's shutter motion is not perfectly even! It's even enough for ordinary photography, but not for this. I have to go to 1/400 or slower (preferably 1/20 or slower) if I want a good flat field, and I have to make sure I'm photographing something that is continuously illuminated, not, for example, a fluorescent light that is flickering at a high frequency.

(9) In Deep Sky Stacker, if you are using flats and flat darks but not bias frames (minimum-length darks), then you should uncheck "Set the black point to 0" under Raw DDP settings. I made the mistake of having this box checked, and my flat-field corrections were quirky — losing contrast at the edges, showing color casts, and so forth.

How do you judge the quality of a flat field image? One quick way is to open it up in Photoshop. It should look like an expanse of featureless light gray. Then use Ctrl-L to stretch the contrast. (Not "Auto Tone," which will change the color.) You will see, greatly exaggerated, the actual information that it records — typically edge darkening, maybe a few blotches, and a faint tartan-like pattern from sensor unevenness. You should not see strong colors or other unexplained unevenness. For example:

Here (stretched) is a good flat field taken at 1/400 second:

This one only shows edge darkening. You might also see blotches corresponding to dust on the sensor. That's why flats are to useful — they correct for the effect of that same dust in the actual picture.

Here (again stretched) is a flat field taken at 1/8000 second and showing shutter unevenness or perhaps a strobe effect from high-frequency flicker of the light source:

(No shutter should be expected to be perfectly even at its highest speed. Remember, you're looking at tremendously stretched contrast.)



Data-losing bug in Voice Record Pro 1.7 on iPhone 5 (iOS 7.0.2)

(has bug)

(does not have bug)

Some of my logs from the last two astrophotography sessions (October 1 and October 5) were lost because of a bug in Voice Record Pro, which I was using to record my notes on the iPhone. The built in Voice Memos app does not have this problem, so it's what I'll use from now on.

If you pause a recording, and then put the iPhone to sleep, wake it up again, and resume recording, Voice Record Pro will not actually add to the recording, even though it appears to be recording properly.


Bright and dark nebulae near Mu Cephei

To the lower left of center in the pictures below is Mu Cephei, the "Garnet Star." Below it you see a large round nebula, a faint cloud of glowing hydrogen. Elsewhere in the picture are dark nebulae in front of star clouds. These nebulae are not very well known even though they are often high in the North American sky.

Mu Cephei

[Reprocessed 2013 October 12.]

This is a stack of nine 30-second exposures at ISO 3200 with a Canon 60Da, with a Sigma 105-mm f/2.8 lens wide open on an iOptron SkyTracker, calibrated against dark frames, flats, and flat darks using DeepSkyStacker.

My equipment on the field

You may be wondering how much equipment I bring with me in order to take pictures like this. Here's what I took to the Peach State Star Gaze. There's a gray piece of cloth over the camera to protect it from dust and dew when it's not in use. What you see here is a tripod, a battery, an iOptron SkyTracker, a camera with a telephoto lens, and a Kendrick dew heater with a homebuilt controller of my own design (a 555 PWM circuit).


Clouds of stars and clouds of dust

Here is one of the pictures I took at Deerlick on October 1. You're looking at the star Altair, clouds of stars near the center of our galaxy, and clouds of dust (dark nebulae) in front of them. This is the same field I photographed from in town last month. The dark country sky helped!

Barnard's E

[Reprocessed 2013 October 12.]

The most prominent dark nebulosity is called Barnard's E. Stack of ten 30-second exposures at ISO 3200 with a Canon 60Da, with a Sigma 105-mm f/2.8 lens wide open on an iOptron SkyTracker, calibrated against dark frames, flats, and flat darks using DeepSkyStacker.


The sky is still there

On the evening of October 1, I had a successful trip to Deerlick and took some very good astrophotos, which you'll see here soon.

To put it in perspective, this was my first observing trip to Deerlick, or indeed to any dark-sky site, since November 2011 [corrected]. It's good to know the sky is still up there!

Also, I had a lot of new technology to play with. I used my iPhone as a voice recorder (for the logbook), as a star atlas, and for e-mail, Web surfing, and telephony. Instead of a telescope, I carried the iOptron SkyTracker. For that matter, this was actually the first time my Canon 60Da had been to a dark-sky site!

And besides some digital work, I took my last Ektachrome astrophotos, using my last roll of Ektachrome E100G, which is in a Canon EOS 10S camera body — the one that can take long exposures without running down the battery; the camera I should have had when I was spending money on a Nikon F3 system just before the demise of film.

That required some kluging. Dew was a problem, so I used masking tape to secure the focusing ring of the lens, then wrapped a 3-inch Kendrick dew heater around it. We'll see what happened... The last film astrophoto was a 25-minute exposure of northern Cygnus, and I'll have it developed in a few weeks and post it here.

The opposite of progress is Congress...?


If con is the opposite of pro, then the opposite of progress is...

...the legislative body that is mismanaging our federal government and has had the entire government shut down for three days, delaying the start of one of my important (funded!) research projects.

If, as a UGA administrator, I had failed to submit a budget on time for anything that I was in charge of, I would have been fired. Congress deserves no less. I wish we had a British-style system where a deadlock in Congress would automatically trigger an election.

I know Obamacare is bitterly controversial. If you were reading this blog when it passed, you know that I also think it is seriously flawed and off-target. But we don't know anything about it now that we didn't know six months ago. Today's shutdown is a totally artificial emergency — not a response to a sudden new development.

Whether Congress does its job properly is entirely separate from whether you're for or against Obamacare. There are calamities that might justify shutting down the normal functions of government, such as, perhaps, preventing sudden entry into an unjustified war. But blocking a piece of legislation that Congress already passed, and that about half the American people support, is not that kind of calamity. It is the job of Congress to negotiate and pass a budget, on time, regardless of how divided the country may be.

People are trying to manipulate you with rumors


And the rumor mill is starting up. People are asking you and me to spread, on Facebook and other media, horror stories of how badly Obamacare has treated someone — tens of thousands of dollars for insurance — thousands in fines — hopelessness and despair — but we don't know exactly who this happened to, or where, or when. This morning's juicy story says that someone was hit with a heavy fine by e-mail during the government shutdown. I think somebody might be making things up.

Verify, ignore, or be made a fool of — those are your choices.

As you know, I am not a supporter of Obamacare; something needed to be done, but not quite that. Nonetheless, Obamacare did pass, and it does address some problems. One of the worst things about hoaxes and rumors is that they divert attention from real, practical ways in which our health insurance system could be improved.

Malicious gossip is a sore spot with me. I never engage in it myself, and as some of you know, I tend to snap at people who try to involve me in it, just the way I would snap at people trying to involve me in theft or adultery. The last time I looked, "Thou shalt not bear false witness against thy neighbor" was still one of the Ten Commandments. It's wrong to spread rumors of unknown truthfulness even if they help your favorite political cause. I know that, sometimes, an unproven political allegation might be of legitimate national interest. But even then, you have a duty to be honest about the evidence (and ignore the story if there is none), not just swallow everything someone said.

If what you are looking for is not here, please look at previous months.