Why stack multiple images, and how many is enough?

During one of my recent projects, the Heart Nebula, I decided I would stack some images at different intervals and report here on my results. I used my SII filter for this test. SII usually will not have anywhere near the signal you'd get from the hydrogen alpha filter. As a matter of fact, I did not use any noise reduction at all on my hydrogen alpha data of this region. I thought SII would be a good filter to use for this test, because we can better see the awful noise and artifacts.

So first of all, why stack multiple images? Stacking images can actually be used in any setting that isnt changing to increase resolution and decrease noise. Or in other words, it will increase your signal-to-noise ratio (SNR) and increases the dynamic range. Astrophotography happens to be perfectly suited in this manner. Deep space objects (DSO) are virtually permanent and remain static for long periods of time.

You also want to stack multiple images because each exposure will inherently have cosmic rays, satellite trails, airplanes, and noise in both hot and cold pixels that stacking algorithms like sigma clipping will correct.

Lets see how stacking more and more images improves the final data we have to work with. Each of these are a stack that has had a Screen Transfer Function (STF) applied in PixInsight so that we can actually see what we have. Without the STF applied, the images are very dark and we would not be able to tell anything. You can see how the images gradually increase in SNR and ultimately become much easier to work with.

Each of these stacks are from images that have been binned 2X2 and 2X drizzle has been applied in the stacking to gain back most of the resolution. Please take note of the black specks (cold pixels) in each image and how they are gradually reduced and ultimately eliminated. These are center crops at 100% and each image stacked is a 900 second exposure.

5 subs

9 subs

12 subs

15 subs

23 subs

28 subs

40 subs

So, how many would be considered good enough to you? Some would say 15-20 subs would be enough and would not want to spend 10 hours on one filter. This is perfectly fine. As a matter of fact, I'm working on one right now that I will likely call good enough after 20 subs with the SII and OIII filters. But you can clearly see, if you look close enough, that adding more and more can only help the final image. There is, however, a point of diminished returns. I feel that 40 works well for me when binning and drizzling the SII and OIII data. I shoot my H-alpha at full resolution (1X1) using 20 minute subs and find 8-12 hrs works very well. When shooting this long I sometimes do not even need to use noise reduction.

The above center crops have had no processing at all. Here is a crop of the final SII after processing.

Final crop of the SII

Here is a center crop of the final image using the hubble palette to assign the color channels. SII for red, Ha for green, and OIII for blue.

Final SHO, center crop

And finally, here is the final full image. The full screen image can be seen here, and by clicking on the icon at top right it can be expanded out to full resolution.

I hope this has shown some beginner and intermediate astrophotogaphers why it is important to get enough subs to make a nice image. It is ultimately left to the astro imager to decide for themselves how many is enough, and there is no right or wrong way to do this.

Happy imaging everyone, and happy Thanksgiving!

Creating Hubble Palette Images

I usually try to post here at the end of every month, but neglected to do so in August. I think the TS 107mm got more than just a little attention, it got a thorough workout: eleven hubble palette images and one HaOIIILRGB image since mid July, not to mention a few more that I yet to process. I have a few on the back burner that I'm saving for a rainy day.

The recent images from the 107mm scope can be seen here, starting with the Wizard nebula and ending with the Cave nebula. I have switched back to the Star71 for some wide field imaging.

I also got quite the workout with the stretch of clear weather. I had a severe case of astrophotographer fatigue, so I let August slip by without a new post. I did, however, get in a lot of practice on creating hubble palette images, and thought I would share a little bit of what I do to create these images here at Little Piney Observatory this month.

Hubble palette images are basically created by using data from three narrowband filters, SII (sulfur) is assigned to the red channel, Ha (hydrogen alpha) to the green, and OIII (oxygen) to the blue. Some people also call this SHO, substituting the SII (S), Ha (H), and OIII (O) in the channels represented as RGB, or red, green, and blue. Just simply putting the image together in this fashion will usually result in an image that is very green, since Ha is almost always very dominant in every image. The process I use to create the image is sometimes referred to as modified hubble palette, because it uses some selective color adjustments to manipulate the colors. However, this "modified" hubble palette is what creates what most are so used to seeing in the hubble images with the classic golds and blues.

To demonstrate, I'll use some data from an image I'm working on right now. I'm back to using the Star71 again to try to do some hubble palette images of some widefield targets. Right now I'm working on IC 1848, the Soul nebula.

I process each image before combining into a color image, being careful to try to keep star sizes as equal as possible in each image. Here are the three channels I will be working with.

SII - for the red channel

Ha - for the green channel

OIII - for the blue channel

As I mentioned earlier, a straight combine using no selective color adjustments tends to usually come out very green because of the dominant Ha channel. I put one together here to show the effect.

I don't particularly care for the green look myself. There are a lot of steps that go into creating the traditional look of the hubble palette images, but it's worth it in my opinion. I will go through the steps here and then show you a much easier and quicker way in the end.

Start out by taking this image with the straight combine of SII for red, Ha for green, and OIII for blue, and using Photoshop, make these Selective Color adjustments to bring out the classic gold colors seen in many hubble palette images.

First, align your color channels by using Levels and looking at your histogram to align the three channels.

After aligning your three channels, stay in Levels and select your red channel and set your black point to 4 and leave your white point at 255. (You will leave the white point at 255 on all channels)

Select your green channel and move the black point to 9 and set gamma (center slider) to 1.27

Select your blue channel and move black point to 8 and set gamma to 1.38

Now go to Selective Color, and with Method set to Relative, choose green from the dropdown menu and set Cyan to -100% and Magenta to -25%

Choose Yellow, Relative mode, and set Cyan to -100% and Magenta to 25%

Choose Cyan, Relative mode, and set Cyan to -1% and Yellow to -100%

Again with Cyan in Relative mode, set Magenta to -25% and Yellow to -100%

The final adjustment in Selective Color uses Absolute as the mode. Select the color Yellow and adjust Magenta to 15%, then select Blue and adjust Magenta to -20%.

Quite a few adjustments there huh? The easy way here is to buy Annie's Astro Actions from Anna Morris. This is a great action set for Photoshop for only $15, and it includes the Hubble Creation action as outlined above, plus a bunch of other very useful actions for processing your images.

Once I run this action, I get an image exactly like the one above, and then am prompted to adjust levels to even out my colors as necessary. Once doing that and clicking "OK" I have an image like this.

Now we have an image with the classic golds we usually see in this type of image, but there are a few more steps. If you do a search of hubble palette you will likely see a lot of these images that have pink/magenta colored stars. This is where I go back to PixInsight and use Pixel Math and type in this expression.

R/K: $T
    G:  m = min($T[0], $T[2]); iif(m>$T[1], m, $T)
    B: $T

symbols: m

Then just drag the triangle over to the image and the pink/magenta is magically gone from all the stars. The pink stars always gave me fits, and I'd like to thank David Ault for giving me this pixel math expression. It has been a tremendous help. Thanks David!! :)

After cleaning up the stars, I go back into Photoshop to work on the colors a little more. I also have Carboni's Photoshop actions and like to use his "Increase Star Color" action to get a little bit of star color here. Its tough to get good star color in narrowband images without taking some RGB data and overlaying that for star color, but this action does help. Be careful though not to over-do it.

Another action of his that I use a lot is his "Deep Space Noise Reduction." This noise reduction action masks off the brighter parts of the image so as to not lose any details, and focuses mainly on the background. It basically only affects the darker regions.

I also prefer more blue from my OIII compared to the greenish-blue hue seen here. To fix this, I use Selective Color and work with Cyan. Adjust the Cyan to -30%, Magenta to 30%, and Yellow to -100%. These adjustments vary from image to image so adjust these to taste. From this point I usually make some Vibrance and Saturation adjustments.

With this image I have only completed capturing my OIII data. I still want to double my SII data, and I need a lot more Ha. At this point I only have 6 Ha subs, but the Ha signal is strong, and I managed to produce this with the data I have. This is very close to what my final image of this will look like. It looks fine right now in this small image, but when seen full size right now the "warts" are easily seen. More data will get this cleared up.

One other thing that helps a lot on some images that are weak in SII and/or OIII, is a very mildly stretched hydrogen alpha image. I integrate that into the color image, then come back and use a fully processed Ha image as a luminance layer. Just saturate the color image to taste and then add the Ha/luminance for the detail and blend as usual. Images strong in SII and OIII do not need this done. They blend very well into the Ha. This area of the sky has strong SII and OIII components and did not need this.

I have really enjoyed shooting narrowband lately and putting together these hubble palette images. I always thought that I enjoyed shooting and processing LRGB more, but lately about all I have been doing is working with narrowband and hubble palette. Its very enjoyable to me and I hope that a few people find some of the information in this post useful for their own imaging endeavours.

So on this day of the Autumn Equinox, happy imaging everyone!

The Cool Observatory

This is a follow up on the last couple of posts here at LPO about the observatory upgrades and motorized focusers.

The new insulated roof has helped a lot at keeping down temps inside the observatory. However, it just wasnt quite enough once we felt the full brunt of Summer here. We have started insulating the walls with the same 1" foam we put on the new roof, and have installed an a/c unit in the wall. I just came in from checking on things in the observatory and it is a dry 86 degrees inside the observatory. Much better than the humid 100+ degree temps we saw before the upgrades. I feel much better about keeping my equipment out there now.

As soon as I finish up the insulation I will put some blandex (waferboard) on the walls and paint it a flat black. Should be really nice in there for all seasons when that is done. I dont want things cold inside, just cooler. If it were very cold, everything would fog up this time of year when I rolled the roof off. As long as it stays below 90 degrees inside all the equipment should be fine.

Now for the fun stuff. When I wrote about motorized focusers, I had yet to motorize my widefield scope, the Star71. The large single speed knob of the Star71 is not removable. This forced Leon Palmer at Rigel Systems to add an extension to the motor shaft and run the gears off of the fine focus knob. It might not be ideal, or as clean an install as the one on my TS107, but it seems to do the job just fine. It is nice to have the focusers on all three of my imaging scopes motorized now.

I enjoyed using the Star71 for a while, taking several widefield images, but it was time to start using the TS Photoline 107mm triplet again. I had been using my RC telescope for most of the Spring shooting galaxies, then the Star71 for several of the large Summer nebulae. So it was time to give the 107 some attention.

I have now completed two hubble palette images with the 107mm and started on a couple more. The first one was NGC 7380, the Wizard nebula. For this image I used 12 subs of 900 seconds each of Hydrogen Alpha, 14 subs of 1200 seconds each of OIII, and 12 subs of 1200 seconds of SII, just shy of 12 hours total.

I spent considerably more time on the next image, Sh2-86, or Sharpless 86. It was much tough than I expected and with the low OIII sensitivity of my 8300 camera I probably should have tried binning, or shot even longer than the 30 minute subs I used. Binning would have been easier I think than shooting 45 minute to 1 hour long subs. Since I used the H-alpha as my luminance layer after using a weakly stretched H-alpha to combine for the color, I think binning would have been fine. I'm hoping to buy a new camera soon with the Sony ICX814 sensor. Until then I think I will stick with targets with good OIII signal when shooting color narrowband images.

Sharpless 86 is a very interesting area in the constellation Vulpecula, near M27, the Dumbbell nebula.  The most striking feature is the trunk-like pillar of dust and gas protruding from the east side of the nebula towards the open cluster, NGC 6823 in the west. Remarkable dark globules of gas and dust are also visible in the nebula, much as is seen in the better known Eagle Nebula in Serpens or the Lagoon Nebula in Sagittarius. The stars in the center of NGC 6823 are about 2 million years old and are predominantly "young" bright blue stars.

I have over 20 hours in Sharpless 86 and threw away a few hours as well because of focus problems. I used 23 Hydrogen Alpha subs of 1200 seconds each, 12 OIII subs of 1800 seconds each, and 13 SII subs of 1800 seconds. I would like to come back to this one and shoot some RGB for star color. If I dont have RGB for star color I prefer to just make my stars white instead of the typical magenta color you get in hubble palette versions.

The two works in progress are the Bubble Nebula, aka NGC 7635, Caldwell 11, and Sharpless 162, in Cassiopeia, and the Eastern Veil Nebula in Cygnus. The SII and OIII signal in the Veil is super strong, probably the strongest signal I have ever seen since I began narrowband imaging. Here's an example of the three channels.

Hydrogen Alpha, 20X900 seconds

SII, 15X600 seconds, binned 2X2 

OIII, 18X600 seconds, binned 2X2

The SII and OIII were shot binned 2X2 and then stacked using 2X Drizzle to match the resolution of the Hydrogen Alpha shot at 1X1. I have put together a couple preliminary images using this data. I would also like to add some RGB to this one for star color.

Bi-Color image using the Hydrogen Alpha and OIII from the above data

Hubble Palette version of above data

So far I have only the hydrogen alpha data shot on the Bubble Nebula. I'm hoping to get started on some SII and OIII for it as soon as the Moon become less of a factor. These images will be posted soon to my Astrobin gallery.

Happy imaging everyone and stay cool this Summer!!

Summer Projects

We have been busy this month at Little Piney Observatory. Summer brings a lot of things, some good, and some bad. Part of the good is the beautiful large nebula in the night sky. It was time to put the RC away and get out the Star71 for some widefield imaging. Since the RC telescope was going to be out of service a while, I decided it was time to give the primary mirror a long overdue bath.

The bad that comes with summer is the awful heat and humidity. Last summer was a constant battle to keep the observatory under 100 degrees. We often lost that battle, and it was such a hassle and a constant worry, that we decided to do something about it this year. Little Piney Observatory received a new roof this month, along with some one inch insulating foam with a reflective backing. The super thin sheetmetal on the roof was removed and replaced with some quality metal and a substantial overhang was added to keep things dry inside.

Lets start with cleaning the mirror. There are several of these GSO RC scopes being used, and I thought someone might like to see just how easy it is to remove these primary mirrors. The GSO scopes come with various names including Astro Tech, TPO, Levenhuk, Teleskop Service, and others. I have owned this scope almost two years. It has seen a lot of use in that time, and with the pollen, dust, and the spiders that found their way into the optical tube assembly (OTA), it was time for a cleaning.

The first thing to do is take the top and bottom dovetail screws out. Then there are two other screws on the side of the OTA that need to be removed. After this the entire rear cell with primary mirror can be removed.

Secondary mirror with primary mirror removed

Carbon fiber tube

Rear cell with primary mirror and baffle tube. Seen here after cleaning.

I chose to remove the mirror from the cell before cleaning. To remove the mirror the baffle tube needs to come off. The entire baffle came off for me in one piece. Others have said the small piece at the bottom of the tube remained attached and they needed some small allen wrenches to remove the rest of the baffle tube assembly. Once this is removed the mirror can be lifted right out of the rear cell. For more info on the baffle tube, please see this thread on cloudy nights that goes into more detail on removing this piece.

After removing the mirror, I ran some warm water in the kitchen sink along with just a drop or two of dish soap. I placed the mirror on a towel and let it soak for about half an hour.

I put about half a dozen cotton balls on the mirror and moved them from center of the mirror out to the edge without putting any pressure on the mirror. Discard these cotton balls and repeat. I used my fingers to get off any stubborn pieces and rinsed the mirror with distilled water. After rinsing I used a micro fiber cloth to remove puddles of water and dried it quickly with a blow drier.

Install the mirror back into the cell and reverse procedure to put the scope back together. I am sure the collimation will need a couple of tweaks to get the mirrors lined up again, but I have not checked it yet. It is sitting on the shelf now awaiting that next galaxy or globular. I plan to check and align the optics soon.

RC ready to be collimated

The new roof that I mentioned required a little bit of money and a lot more work. It also significantly increased the weight of the roof and made it a little harder to open and close by hand. However, with the drop in temps inside the observatory during the daytime, it was very much worth the time and effort that went into it. I plan to eventually add a gate opener to the roof, but it's nothing that needs to be done right away.

Removed old sheetmetal from the roof and removed aluminum flashing from the sides

Added new gable ends and solid boards in place of old flashing

Added new 2X4 laths for attaching new sheetmetal

Installed new foam insulation

Installed new sheetmetal and ridge cap

New roof with added overhang = Much improved!!

I still have to find something (lightweight) to cap off the ends of the newly added overhang, but I am very happy with the new roof. Temps inside now get to only 1-2 degrees above outside air temps, and the inside remained dry after a substantial rainfall dropped by a passing thunderstorm.

I'd like to end this month's post by showing off a few of the wonderful summer nebulae that require a short focal length scope to capture (without having to do mosaics). For these I used the William Optics Star71 telescope, a very small f/4.9 apo with a focal length of just 348mm.

My first Image of the Day on Astrobin!!!

For full size views of these and other images, please see my gallery on Astrobin. I just started a couple more projects, including a widefield Sh2-101, the Tulip Nebula in Cygnus. Cygnus is a treasure trove of fascinating targets that will keep me busy for many summer's to come.

Have a safe and happy summer everyone! :)

Motorized Focusers

In my last post, I told you I was dusting off the RC telescope to get ready for galaxy season. I managed a few galaxy shots, but had finally had my fill of the scopes stock focuser. The focuser is by far the weak point of the GSO RC's. I also like to rotate through the RGB filters when getting my color data. This is difficult to do when the focus points of each filter are not the same. I decided to go with the Moonlite CSL 2.5" Large Format focuser with High Resolution Stepper Motor. What a fine piece of equipment this thing is! It is really a work of art.

I chose the M68 threaded drawtube version. It came with 2" and 1.25" threaded ends with compression ring and thumbscrews. I may eventually purchase the M68 to M48 adapter to make the entire image train threaded. It seems to be working just fine like it is though.

The grey piece hanging off the motor is an adapter I got from Rigel Systems for $10. This adapter lets me use my Rigel controller that I purchased a while back along with a stepper motor for my TS 107mm triplet.

This is my second Moonlite focuser that I have bought and I really like them a lot. The Teleskop Service 107mm scope came with a nice 3" rack and pinion focuser so I decided to keep the focuser and just order the Rigel Systems kit for it. Here are a few shots of the Rigel setup on the TS107.

Rigel usb-nStep controller

Both of these work great with Sequence Generator Pro. Once focused you can save the focus position and it will automatically go to this preset position when the filter is changed. It also allows for autofocus, which can be programmed by temperature if you wish. You can also set it to run autofocus upon a filter change, once an hour, or a number of other options. Once autofocus is ran there is a box that can be checked to adjust the focus points of the other filters by the same amount.

The guys at The Astro Imaging Channel have a really good video that goes into this much more than I can here. If anyone is interested it can be found here.

Galaxy season was just getting started when I last posted in March. I have several more since then, all taken with the RC, along with the Globular Cluster M13, and the interesting planetary nebula Jones-Emberson 1. These can all be seen in my Astrobin gallery.

The motorized focusers are really very nice and I'd hate to be without them now. I was reminded of this last night when I got out my William Optics Star71. This scope has a nice 2.5" rack and pinion focuser and it will be receiving a Rigel motor very soon!!! I was able to get in a couple hours of luminance data with it before the clouds moved in. I am very excited about this image and am hoping to get some good skies to finish it soon. This is just 6 subs of 1200 seconds each on NGC 7023, the Iris Nebula, and the surrounding dust in the region. I'll leave you this month with this work in progress. I hope to share the finished image soon! :)