Moonlight Imaging

I don't mean yesterday's blue moon but a pair of Sharpless objects. I've started working on the AL Arp Galaxy program, and it suffers from a seasonal bias--external galaxies tend to be seen best when placed away from the plane of our own galaxy. The Bright nebula program had an opposite bias in that most nebulae are in or close to the plane of the Milky Way. When the Milky way is well placed in the sky you're kept busy imaging nebulae; when it's not, you have almost nothing to do. The opposite holds for galaxies.

Two examples: The BN list has 15 objects in Cygnus, while the Arp has only one; The BN has one object in Ursa major while the Arp has 34!

The way to avoid the "hurry up and wait" problem is to do multiple programs at once, choosing programs that have complementary biases. So I think I'll also do the Planetary Nebula program (concentrated in the galactic plane) along with the Arp (concentrated at the poles).

Here's an image for the PN Program. It's not my first; for that I'll have to look back through my old images. This one is Sh 2-71, very bright and sitting just west of a much larger and dimmer Sh 2-72.

Sh 2-71 at right, the much more extended and dimmer Sh 2-72 at center left.

This image is in H-alpha and is another illustration of the power of narrowband imaging. These objects were in Aquila and the full moon was in Aquarius, not very far away. The moon was up for the entire data acquisition time and I was imaging under a red zone sky as well. In other words, the sky was bright.

(Imaged using a TV-102, ST-8300M; 6 x 600s autoguided exposures.)

Last night the unbeatable factor was clouds; a lovely clear night slowly went overcast as I was collecting O-III data. I had to settle for an hour of H-alpha.

It's nice to know that even though we have no remedy for clouds, we can image deep despite the moon.

Finally!

This is the first image of the year! Let's just say it's been a bad year. Amazingly everything worked smoothly, including me. This is yet another AL Bright nebular object, LBN 331.

A so-so picture, but I'll take it--LBN 331 is a brightness 4 object, so given my skies I don't expect much from 1.6 hours total exposure.

LBN 331

Details are here.

Tonight's supposed to be clear again (we'll see), so with luck I can bag another object, or go back to this one and shoot some RGB so that the stars have color.

Are Baader filters Parfocal?

At a recent star party I made an offhand statement that I doubted that my narrowband filters were parfocal. That set me wondering if I was correct and how would I go about determining this for all my filters. So I did a little experimenting one night by imaging Betelgeuse through all my filters.

I decided to set focus with the luminosity filter in place, and then cycle through all my filters seeing if the focus changed. Sorry, but I don't have a focuser with a digital readout. To get some sense of any focal shift I put a Bahtinov mask on my scope and imaged the results. I used three stars of different spectral types (O, A, and M), just in case the results would be influenced by wavelength. (I was imaging through my TV-102, so I didn't expect to see any significant wavelength dependence.) Here is a test strip for Betelgeuse (click to see full scale image):

Betelgeuse imaged through various filter configurations with no adjustment of focus

I used 5 and 25 s exposures for the broadband and narrowband filters, respectively, and processed only using ImagesPlus digital development. It looks like the central spike produced by the mask moves little if at all as the filter is changed. As expected, the focus shifts significantly when changing to no filter in use. So how much shift is significant? Here is an image resulting from 1/12th turn of my fine-focus knob away from focus:

Appearance when defocused by about 0.3mm

This is approximately an adjustment of 0.3 mm in focus and the central Bahtinov spike has shifted quite a bit. From past imaging experience, I think my final focusing "tweaks" are about 1/5 of this in magnitude.

So are the filters (Baader LRGB and 7 to 8.5 nm Ha/S-II/O-III) parfocal? They look close to being so. To my surprise, the broadband and narrowband filter sets seem to be almost, but not quite, parfocal with each other. The results were the same for the other two stars, Sirius and Altinak. But I'll probably still refocus every time I change filters in the future, particularly when using anything less than an apochromatic telescope or there's been a change in temperature.

Photoshop Actions

I spent a little (emphasis on little) of my holiday money on the purchase of two Photoshop action packages for astrophotography. My simple goal for today was using the packages to reprocess one of my old images. Here's my original take on NGC 6781, a small planetary in Aquila. This was shot in June 2012 using H alpha and OIII narrowband filters with my ST8300M CCD and C925. This was my original "final" image:

Original Processing

I used the OIII data for both the blue and green channels, which produced a washed-out cyan in the nebula's interior. Other defects in this image include some uncorrected mild vignetting, a noisy background and almost colorless stars. Using the actions it was an easy task to generate a synthetic green channel, clean up the background, give the stars some color and pretty much obliterate the vignetting (without any flats). Wow! That's not bad for a day's work with new software. Here's the result:

Processed using actions

Not that it's perfect by any means. I think the background is too black and I've lost a lot of the dimmer stars. I'll take it as good enough for the first day of using the actions.

The two packages are Astronomy Tools ($21) and Annie's Astro Actions ($15).  There's some overlap between the packages, but they're so inexpensive that you can get both!

Recent Imaging

I've been imaging over the past few days. The horsehead image from the last post did get its green and blue parts taken:

Flame and Horsehead nebulae

The Moon was Gibbous and nearby; some of its light snuck in and caused the blue-green flare at lower right. This has since been solved by extending the dew shield with some craft foam.

I also had a couple of clear nights during which to image the Rosette nebula. I collectd H-alpha the first night, but it was partly cloudy the second night and I didn't get very many OIII frames.  Consider this a preliminary composite image:

Rosette Nebular (NGC 2244)

The Moon was even brighter and closer to the target than for the previous image, but the dew shield extender worked fine. Next chance I get I will collect more OIII data, and for the first time try to get some SII as well.  I'll get a Hubble Palette image yet!

Narrowband Imaging V: Image Processing

There's not much I can write here that hasn't been written better by others with vastly more experience than I have. Here are a few observations:

• Calibration, aligning and stacking of the collected light frames is important, but the processing that follows is where the art enters in. Most software packages dedicated to astrophotography will do the calibrations, alignment and stacking in an automated fashion. All you need to do is set a few preferences and wait.
• Whatever software you use for postprocessing, try to become familiar with its many features.
• Experimentation is good. Try to jot down the steps you take when in the final stages of postprocessing. This will help you replicate your methodology.
• Don't delete your original images.  As you learn more you may wish to reprocess them.
• These are your images.  You're the ultimate audience, and the only person you're trying to please.

When posting images for others to view:

• Don't post large full-scale images to forums. Reduce your images to a maximum dimension of about 800 pixels. Provide a link to a full-scale image housed on a web-based photo service if you want to people to see it. 
• When posting .jpg images, don't overdo the compression.  Overcompressing produces distracting artifacts, particularly around stars.
• Placing copyright text on an image is basically a waste of time. The only way to prevent people from using your images without your permission is to not post them. (This is a personal thing with me; I think copyright notices look pretentious.)

Speaking of matters of personal taste; here are some of mine:

• I don't like backgrounds that are absolutely black; a dark neutral gray is perfectly fine and more realistic.
• Stretching an image is easy to overdo and can cause hard edges on nebulae. I usually prefer a softer-looking image that is a bit understretched.
• If you want diffraction spikes on your stars, don't use software to create them. Get a telescope with a spider-supported secondary, or create a mask that holds fine threads across the optical path.  Better yet, learn to like images without spikes. 
• Don't overdo filtering. Unsharp masks and other sharpening algorithms can introduce artifacts that detract from image quality. Toggle between before and after to watch what an enhancement does to nebular texture. If it introduces features that are barely there in the "before" image weaken the enhancement.

Narrowband Imaging IV: Setting Up

To give you an idea of what is involved in performing NI, here's what an NI newbie (me) does to set up for an imaging evening.

During dusk, haul things out of the house and do other simple tasks:

1. Carry out Mount (put it on the pier, check level. I have a cement pier, which spares me having to lug out the field tripod.)
2. Carry out Work table (unfold and set up)
3. Carry out Counterweight and counterweight shaft (attach to mount)
4. Carry out Telescope assembly (put it on the mount)
5. Carry out Two deep cycle batteries (one for laptop and dew preventers, the other for the mount and CCD. I choose to run from batteries in preparation for setting up at remote sites without AC power.)
6. Carry out Tackle box with cables, etc.
7. connect Autoguider with Mount
8. connect Mount to Battery
9. Temporarily put CCD on telescope
10. Balance the load
11. Replace CCD with diagonal and illuminated eyepiece
12. If using dew preventers, put them on the telescopes and connect to battery

When Polaris becomes visible:

1. Adjust mount to bring Polaris into view in Polar Axis bore.
2. Align finder with imaging scope
3. Turn on mount and do two-star align, augmented with two more calibration stars
4. Do software Polar align (If I were patient, I would do a drift align here; at this point I should repeat the two-star align with calibration stars, but usually I don't--goto is reasonably accurate.)
5. Turn on dew preventers (if needed)
6. Connect laptop to battery and turn on
7. Swap out diagonal and illuminated eyepiece used during alignment for CCD
8. Connect CCD to battery
9. Connect CCD to laptop
10. Start software for CCD control, turn on cooling
11. Slew to bright star (be careful of CCD-to-laptop cable)
12. Connect Autoguider to laptop
13. Focus CCD with narrowband filter
14. Slew to target (be careful of CCD-to-laptop cable)
15. Compose the image frame using luminosity filter (do NOT refocus)
16. Connect Autoguider to laptop
17. Start PHD
18. Select guide star and calibrate PHD
19. While PHD is calibrating, switch to narrowband filter and expose a test frame for desired exposure time
20. Check that CCD cooling is correct.
21. If test frame is OK (composition is correct, etc.) aside from star streaking caused by PHD calibration and PHD is done and locked on the guide star, begin shooting light frames.
22. Periodically recheck progress, watching for any irregularities in the latest light frame. Also check PHD for uncorrected drift

When changing filters, with new filter in place repeat step 13 and resume shooting light frames, periodically rechecking the system as in step 22.

If you don't have dark and bias frames for this temperature and exposure time, shoot them at the end of the evening. If I shoot flat frames, it's usually the next day. So far I've not seen the need to shoot sky flats. I would were I using LRGB filters.

When all imaging is completed:

1. Return mount to home position
2. Power off mount
3. Turn off CCD cooling and disconnect from software
4. Disconnect CCD and Autoguider from laptop
5. Disconnect Autoguider from mount
6. Shut down laptop, disconnect from battery
7. Carry in laptop
8. Disconnect CCD from battery
9. Stow cables in tackle box and carry in
10. Carry in batteries and telescope assembly
11. Carry in telescope assembly
12. If the weather will be OK, leave the mount on the pier and cover it.
13. If bad weather is on the way, remove the counterweight and counterweight shaft and carry in. Take down the work table and carry in 

It all looks so simple this way. I'm sure I've left out some steps, but I'll add them as they come to mind.

The most difficult part for me is the two-star alignment with calibration stars followed by polar alignment, because I can't see the western part of the sky.

I also tend to spend quite a bit of time on image composition, trying to get it "just right."

Also time-consuming is focusing because of how stars are dimmed by a narrowband filter.  I have tried using Jupiter for focusing to good effect. I've also used a Bahtinov mask a few tunes, but so far it doesn't save that much time or improve image quality any over what I can do with my camera control software's focus utility. I need to work with the mask more.

If you have an observatory you can pretty much skip the entire "During dusk..." section.  That's where I hope to be by the end of next summer!

Next Time: Image Processing

Narrowband Imaging Esthetics

I left out an important distinction between NI and one-shot color imaging: What the end product looks like. DSLRs and one-shot color cameras are very good at reproducing the actual colors of an object. Reflection nebula and galaxies shine by scattered or direct starlight. Both of these are broadband, and so not well captured by NI. NI maps emission bands to the RGB portions of the final image, resulting in colors that are far from realistic. (Think of Hubble nebular pictures, with their bright greens and blues.) This is why you often see hybrid images of galaxies taken through H-alpha and RGB filters. The former accentuates the emission nebulae within the galaxy, while the broadband filters reveal the stellar portion of the galaxy.

The upshot of this is that if you expect realistic colors, NI may not be for you, or you should resign yourself to occasional trips to dark sky sites for some objects. If monochrome images or false colors don't bother you, then NI will deliver what you want from your own urban backyard

Narrowband Imaging III: Guiding

Narrowband filters pass light only near discrete wavelengths. This makes them effective at reducing light pollution, which is mainly spread across a broad range of wavelengths. The downside is that faint nebula don't generate all that much radiation even at the passed wavelengths.  This means that you will probably want to have a total exposure times of several hours using each filter. (Narrowband imaging is not for those who need instant gratification!) You can do this by combining a large number of short exposures or a small number of long exposures.  Generally, the latter is preferred.

Let's assume you want to use modestly long five minute exposures (also known as subs).  It's not unusual for skilled imagers to use exposures as long as 20 minutes to an hour.  There are a number of things that can happen during five minutes to degrade your images; wind, poor polar alignment, and imperfections in your mount's mechanics can all lead to misshapen stars instead of nice round ones.

Your first line of defense is polar alignment. A good but probably not adequate alignment can be obtained using a polar axis scope. A better alignments results from using a software method that is built into many modern mounts.  The best alignment results from using the declination drift method. The drift and software methods require you to use an eyepiece with illuminated cross hairs. The drift method demands time and patience, and can be confusing the first few times you try it. But it's the only real way to get a nearly perfect alignment. With a near-perfect polar alignment you can take exposures of five to ten minutes with minimal defects. The longer the focal length of your imaging telescope the greater the need for an accurate polar alignment.

A great polar alignment virtually eliminates field rotation when using a GEM mount or wedge-mounted fork. There's still a problem of small motions in right ascension and declination caused by imperfections in your mount's mechanics. These can result in distorted star shapes. There are several ways of dealing with this: autoguiding, periodic error correction (PEC), and adaptive optics. The latter remains very expensive at this time, so let's concentrate on the first two. Autoguiding uses a separate telescope and camera to track the motion of a guide star. As mount imperfections and wind cause the guide star to move, software sends pointing corrections to the mount.  Used alone, autoguiding does a good job of eliminating drift.  When used with PEC, autoguiding reduces tracking errors even more.

Some imaging CCD cameras come with a built-in autoguiding sensor and use a guide star from the edge of the field of view. In some models this places the autoguider sensor behind the narrowband filter, which is a disadvantage. Many people prefer a separate autoguiding camera and telescope  for this reason, and for the reduced cost.

Thanks to modern software, your autoguiding telescope's focal length can be much shorter than that of your imaging telescope.  A favorite autoguiding telescope is the ubiquitous f/5 80mm achromat. These are generally easy to buy used. The autoguiding camera can be something designed specially for guiding, or an older CCD camera such as a Meade DSI.

If you decide to use an autoguiding telescope, you must mount it securely.  Flexure is what happens when the orientations of the optical axes of the guiding and imaging telescopes change relative to each other. Usually this happens when the mounting hardware bends or shifts. You don't need the two optical axes parallel, but you do need to keep them from changing their relative orientations.

The most popular software for autoguiding is called Push Here, Dummy (PHD). It's freeware and excellent. PHD is simple to use, but it has some parameters you can adjust to make it work better with your setup. Experimentation is encouraged.

Some autoguiders incorporate the guiding software within them. This convenience raises their price.

None of the above methods will do you much good if you don't balance your imaging system properly. it's important to balance the mount's load around both the polar and declination axes. In my experience, the nights I've had  the most trouble are those when I didn't do a good job of balancing. Usually I balance with the imaging telescope in the orientation it will have during imaging.

Next Time: Setting up

Narrowband Imaging II: Components

Last time I wrote about reasons for why you might want to give narrowband imaging a try.  To summarize, try it out if you enjoy astrophotography, live under significant light pollution, and enjoy a challenge to your learning skills and budget.

Here's what you might need in addition to what you already have:

• A monchrome CCD camera (includes adapter for prime focus imaging, power supply, software for camera control)
• A laptop computer
• A mount that is capable of carrying the weight of your telescope and camera
• Narrowband filters for Hydrogen-alpha, OIII and SII wavelengths
• A filter wheel

Let's consider each in turn, in each case looking.

CCD cameras come in a bewildering variety.  If you've been using a DSLR with a nice big APS-C multi-megapixel sensor, don't automatically assume bigger is better. A one megapixel CCD is really all you need for imaging.  Smaller sensors don't extend as far off axis where distortions such as field flatness and coma can become apparent. The also make smaller images that are less demanding in terms of memory, download time, and processing time. Smaller CCDs also allow you to use 1.25" filters, which are much less expensive than 36 mm or 2" filters. The downsides of smaller sensors are that their field of view is smaller and target acquisition and composing can be more difficult..

If you can afford it, get a CCD camera with temperature regulation. An example is the ATIK 314L+. An unregulated CCD is less expensive, but you'll probably want to shoot dark frames before and after each imaging session. (Some CCDs are said to generate so little thermal noise that dark frames are unnecessary. My opinion is that dark frames are always worth using.)

Options: Most CCD cameras come with an adapter, either 2" or 1.25" for use in a telescope focuser. This allows you to do prime focus imaging. Most manufaturers either include an AC power supply or sell one separately. Unless you plan on running off a battery, be sure to get a power supply.   Software to control your camera during imaging is usually included with the camera, but ease of use may vary.  If you're not happy with the included software you can opt for a commercial product or choose freeware for this task.

A laptop computer is the central nervous system of imaging.  Your camera control software will run on it, passing instructions to the camera via USB cable. The acquired images will be sent to the laptop for storage by the same cable. After gathering images you may use the laptop for processing.  Processing images can be time comsuming, so if you need to buy a laptop make sure it has enough speed and resources.  Software for Windows laptops is a little more common than for Macs. Avoid any version of Windows released prior to XP. If you want to use a Mac, check first for software supports it, and confirm that software will support the CCD you intend to purchase.

The mount is the most important component of imaging. Add up the weights of your telescope, camera, filter wheel, and autoguiding telescope (if any) to see how heavy a mount you need. It's typical for the weight-carrying ability of mounts to be described for visual observing. Imaging is more demanding.  Generally speaking—and this is very general—any mount that sells for less than $1,000 can carry imaging gear equal to about half of its stated carrying ability. More expensive mounts have carrying weights that are closer to those suited for imaging. If in doubt, consult the Cloudy Nights forums for the opinions of actual users. Even if you don't intend to use autoguiding, I would strongly suggest obtaining a mount that accepts autoguiding commands.

Narrowband filters are a source of continual debate. Are inexpensive ones as good as those that are priced dearly?  Again, Cloudy Nights forums can be a useful guide.

Filter wheels are more than a convenience because they help keep filters clean while in use. If you already have a filter wheel loaded with LRGB filters, you may wish to obtain a new carousel for swapping your NI filters in and out. Consider an 8-filter wheel, so that you can house both narrowband and broadband filters in it, in the event you go to a dark sky location for imaging. Some filter wheels can be controlled by commands passed by USB.  This is a very convenient feature well worth considering.

Imaging Telescope. This isn't all that important. All you need is an achromat with a good focuser. NI does not require an apo, as chromatic abberation is largely eliminated by using narrowband filters.  (Remember to refocus when changing filters!) You may want to upgrade your focuser if it's not two-speed, it slips when carrying the weight of your camera, or it won't keep the camera squared up to the optical axis. You can also use a camera lens as a telescope, but that can get into messy issues of adapters and spacing rings. (I plan on doing this for wide field imaging in 2012.)

My personal NI setup is what might be considered an example of an high entry-level system: If it seems a little pricey, it's because I was sure I would take to NI and be doing it for years.

Camera: SBIG ST-8300M (yes, I went for the megapixels and big sensor size; if you are more frugal than I am, do as I say, not as I do)
Mount: Celestron CGEM (my total imaging weight is about 20 pounds) There are a number of mounts in this price range, and all get good reviews
Narrowband filters: Baader 7nm (not the cheapest, but certainly not the most expensive.  They have worked well for me)
Filter Wheel: SBIG FW8-8300 (USB controlled)

If you really get into NI, you can easily spend many thousands of dollars on premium equipment. You have been warned!


Next Time: Guiding during long exposures.

Narrowband Imaging I: Motivation

(First of a series)

Disclaimer: I do not presume to be anything other than a beginner at narrowband imaging.  All I've done so far is spend one summer and fall getting started. So what I present here should be taken as one person's opinions based upon reading and limited experience.

Why get into CCD narrowband imaging (NI)?

• You've been DSLR or one-shot color or broadband CCD imaging, but have become dissatisfied with the limits it imposes
• You live under light-polluted skies and don't want to confine your imaging to trips to dark sky sites.
• You want a challenge.
• You have some disposable income.

DSLR cameras are relatively easy to use and provide high-resolution images. An amazing amount of hardware and freeware exists for using them in astrophotography. When modified they can provide fair sensitivity to the wavelengths of light emitted by many nebulae.

The central problem with DSLRs is that they are not temperature-regulated.  The optical sensor within a DSLR operates at a temperature near the ambient.  The warmer the temperature, the more noise generated by the sensor.  This noise shows up in images as a random graininess that grows brighter with exposure time.  This accumulated noise can be somewhat compensated for by shooting dark frames (exposures where the shutter remains closed) and subtracting them from the light images (the pictures of what you want to image). This compensation is limited because it's hard to match the temperature of the dark frames to the temperatures at which the light frames were exposed.  Ambient temperature usually changes during the course of an evening, so you're forced to create dark frames at a temperature only approximating the temperature at which the light frames were taken. If the dark frames are cooler than the light frames you get undercompensation; if too warm, overcompensation results; in either case you lose some of the detail in your image.

One-shot color CCDs are a common alternative to DSLRs. When equipped with a regulated cooling system they can hold the CCD sensor at a fixed temperature, allowing better use of dark frames. In fact, the regulation allows you to preshoot dark frames so that you are spared the need to shoot them during each imaging session. The light sensitivity of color CCDs is also better than that of DSLRs.

Broadband CCD imaging employs a monochrome sensor that is very sensitive to wavelengths of visible and near-infrared wavelengths. As its name implies, it produces gray-shaded images.  Exposures through red, green and blue filters are combined to produce a color image.

A serious shortcoming of DSLR, one-shot and broadband CCD imaging is how poorly they cope with light pollution. Light pollution is generally a broadband phenomenon, filling the entire visible spectrum with noise (i.e., undesirable brightness). Under reasonably dark skies, all  three methods work well. In a suburban or urban location, light pollution produces a serious decrease in signal-to-noise ratio as it overwhelms the photons you're trying to gather from your target object.

NI addresses this problem by employing special filters that pass light only at the wavelengths at which emission nebulae radiate. Light pollution is largely filtered out, whether it comes from urban glow, neighbor's yard lights, or even the Moon. For nebular targets the urban imager can be just as successful as one at a dark site.

There are prices to be paid for this ability.  Narrowband filters can be expensive. Exposures must be very long, requiring a good mount and guiding.  Image acquisition and processing are more difficult. Non-nebular targets such as star clusters and galaxies generally don't turn out as well. But the fact is that NI offers the urban or suburban imager a wide range of targets that would otherwise be impossible.

Next time: A basic NI setup