Jump to content

  •  

* * * * *

Low Cost Deep Sky Astrophotography


Discuss this article in our forums

Low Cost Deep Sky Astrophotography

 

Back in the 1980s when I was a teenager I was interested mainly in two things, photography and audio. My parents were not exactly poor, but money was just enough for the essentials of life and a budget vacation now and then. My pocket money was thus very limited. While other children for instance owned a turntable an amplifier and large speakers, a Walkman was all I could afford. In photography it was similar. A Kodak Instamatic is not exactly a high end camera. Things did not change much when I was a student. So for roughly two decades I was interested in a few things that really thrilled me but the lack of money always stood in the way. Now that my 50th birthday is near things have changed for the better and I was able to buy what in my eyes is a decent astro photography rig. Many CNers are in the same shoes now as I was then, especially in the current COVID crisis when salary is reduced or people even lose their jobs. Having plenty of time it is, of course, the right thing to do something and not just hang around. Being thrilled about a topic obviously causes people to accept a lot of compromises. In this article I would like to show possible ways to get into this hobby even with a very low budget. More importantly I would like to show what limitations come with low priced equipment and hopefully save someone from great disappointment when the first image is processed. I will try to present as many images as possible because an image with weird stars and a blurred galaxy has more truth in it than a product description like "a lens well suited for all situations".

 

Disclaimer

I am not associated with any of the brands named here nor do I receive any payment or free trial products from anyone. All items introduced here just happen to be the ones I own.

Text in italics is to represent typical Cloudy Night content that is not exactly copied from existing threads but appears now and then in a similar form. The acronym AP will be used for AstroPhotography and OP for Original Poster from now on. There will be some math but I try to keep it simple. Prices are for a basic orientation only, given in USD and picked from major internet stores including used parts that vary in price and quality.

As this article is probably interesting for absolute beginners I will explain terms that are well known to astronomers but may be foreign to a newbie.

 

The Zero Invest Plan

Hi folks, I've got a with a kit lens and a tripod and would like to start AP. Would 30 seconds be a good exposure time or do I need more? Please help!

 

Two important points to discuss here. Let me start with the tripod and interpret the pseudo post to mean that neither a mount nor a tracker is used. The earth rotates while the stars stay in place. It does so surprisingly fast as the following example may show: The sun is not a small object. How long does it take from the point when the lower edge of the disk touches the horizon until it completely disappears? The earth rotates one full turn, 360°, in 23h 56min roughly. For simplicity let's calculate with 24 hours which is 1440 minutes. Divide 1440 by 360 and you will find that the earth rotates one degree in only 4 minutes. The sun is half a degree so the answer is two minutes. Although I know how to calculate it, let's use a search engine to find the field of view (FOV) for a 70mm focal length lens and an APS-C sized sensor. The first results come up with something like 20°x13°. Let's assume the camera has got 6000x4000 pixels. Using the longer side for a quick calculation: 6000 pixels spread over 20° equates to 300 pixels per degree.  When the earth moves 1 degree in 4 minutes (240 seconds) it takes up 1.25 pixels per second with the kit lens zoomed to maximum (70mm). Even when the original image is reduced by a factor of 3 to match the typical screen resolution it takes less than 10 seconds to end up with star trails. Let's ignore for the time being that the size of the object is far too small for the focal length and think about the brightest deep sky object in the northern sky, the Great Orion Nebula. Almost all beginners start with this object and so did I. The following can of course be discussed but let's assume with a fast lens it takes about 10 minutes of total integration time to get an image with a tolerable amount of noise. With an exposure time of 5 seconds this makes 120 single frames that need to be stacked. Being on a budget we assume that a mediocre PC or laptop as available, not the newest overclocked high-end gaming machine. After some struggle with a program like Deep Sky Stacker and a few hours of calculation the new astro photographer shouts "Hurrah!" Here is my first astro image and what I probably would have posted, had I been at CN at the time:

 

The Great Orion Nebula, is a diffuse nebula situated in the Milky Way, being south of Orion's Belt in the constellation of Orion. It is one of the brightest nebulae, and is visible to the naked eye in the night sky. M42 is located at a distance of 1,344 ± 20 light years. (from wikipedia)

 

 

Reply:

Great image, mate. You could crop it a bit and put the Orion Nebula more to the center.

 

OP:

Thank you so much for your suggestion. Here is a crop. I was not sure if it looks more dynamic when I rotate it by 90 degree. What do you think?

 

 

 

I was so thrilled to have taken a photograph of an object in space. Fun aside, there are a couple of things to learn from it. While the stars are pinpoint in the full image the crop shows some trailing. Partly it is the trailing but even more it is the limited resolution of a low-priced lens that makes this 100% crop lack any detail. A decent telescope on a decent mount would show a lot more detail even if the object is only 30x30 pixels in size. So here is the first limitation to the zero invest plan. With a focal length of 35, 50 or 70mm almost all deep sky objects will be less than 100 pixels in size and these pixels do not resolve very much. Limitation will be a word I have to use often in this article.

 

The next thing to learn here is what you do not see: a background. A natural looking image has a gray background with just a tiny bit of noise that makes it look like a photo and not like a piece of plastic. In most images stacked from a huge number of short exposures the background is very noisy and some cameras also add dark or colored bands in that case. You have three choices:

a) Leave the background as it is, don't clip it off. It will look like a green brownish pixel soup.

b) Apply a huge amount of noise reduction. It will now look like a piece of plastic.

c) Clip it. The sky will be pitch black. Like it is in space! Just kidding.

 

The static tripod in fact limits the background to one of these three unsatisfactory choices. We will see option a) in a minute.

What if the full image's caption was "The Constellation of Orion"? The background aside it is perfect now. The message is that the common DSLR and its kit lens plus a static tripod are basically suitable for images of whole constellations and the Milky Way. That's not a bad thing. But when it comes to single deep sky objects the zero invest plan won't work.

The Tracker or Cheap Mount Plan

For gifted tinkerers the search term to use is "barn door tracker". It’s a very simple device with a simple hinge like a barn door, thus the name. Added to it is some sort of a mechanical clock or kitchen timer or even a servo, a bit of fishing line or a threaded rod and the result is something like this:



from www.instructables.com, license CC BY-CN-SA

 

If tinkering is not your strong point, professional motor driven trackers, some even with a polar scope are available for about $300. Even a purely mechanical one is available for less than $200. I do admit I never took a photo using a tracker but nonetheless I dare say that great photos of deep sky objects can be taken with astro trackers. Thousands of photos in the internet clearly demonstrate it. No matter if it is the DIY barn door or a major brand astro tracker, they all come with two basic - here the bad word comes again - limitations:

a) Payload. They are designed to carry a DSLR and a not too large lens, not more.

b) Maximum exposure time. Actually two limitations apply here, one being the drift caused by less than perfect polar alignment, the second being periodic error. The latter is caused by small eccentricities of the gears and makes the tracker be a bit ahead or behind the earth's rotation.

The good news is that now exposure times in the range of one to a few minutes are possible, depending on the payload, and the polar alignment error. That is far better than a few seconds.

 

For a bit more money you can get a real mount including Goto. This one is on sale right now. Or wait until Black Friday.

 

(Goto refers to a computerized mount with some 10,000 object coordinates built in so that you can select an object from a list and the mount places it in the field of view by the push of a button.)

 

While Goto can be a huge help for a beginner the extra money does not always mean that the possible exposure time is much longer. Even if the tracking is better one would of course want to use a longer focal length to have an object a bit larger than 30x30 pixels. With the longer focal length the tracking needs to be more precise. So, with better mechanics you can either take longer exposures with the same focal length or use a longer focal length with the same short exposure time.

 

The tracker is one thing, but what about camera and lens? I made up my mind to do a real-life test. As I do not have a tracker and no longer have the cheap mount, I put a low budget combo piggyback on my telescope while imaging the pinwheel galaxy (Messier 101). The budget camera was my Canon Rebel T3i (600D in Europe), available used on Ebay for $300. I dug out an old Tamron AF 70-300mm 1:4-5.6 LD which is available on Ebay for $100. Convinced of my mounts tracking capabilities I dared to use the lens at its maximum focal length of 300mm. As these lenses tend to slip when pointed straight up, the zoom ring was fixed with some duct tape. For perfect focus I used a small self-made Bahtinov mask, less than $3 for the material, cut out on my girlfriend's die cutter for free. (Using friends' tools and machines is part of the low budget way! Do you know someone who owns a 3D printer?)

 

 

As far as the "serious" imaging with my regular AP rig is concerned I took 120 subs of 5 minutes over three nights which sums up to 10 hours. For the low-cost experiment, I used a mere 4 hours and 24 minutes of data. In order to account for a tracker’s capabilities, I used an exposure time of only one minute for the low-cost setup. This results in 264 sub frames. My laptop is two years old now, a quad core i7 at 2.6 GHz, 16Gigs of Ram and a 250Gig SSD. In my eyes this is not a budget laptop. I did expect it to take a bit longer but I did not expect what followed the next morning: I ran out of disk space twice during processing because all the single steps (image calibration, cosmetic correction, dEbayer and registration) had been saved to disk. So, I shoveled all images that were no longer needed to an external hard disk. Overall, the pre-processing took me about 6 hours. When stacking I ran out of RAM and swapping slowed the laptop down so much that even the mouse pointer no longer moved. So, one big problem of budget imaging is the limitation to short exposure times. That in turn means many sub frames and that in turn means that a lot of RAM, disk space and computing power is needed. That contradicts the low budget plan. The basic principle is often observed. Either save the money and invest a lot of time or save the time and invest a lot of money. There is no need to stack all sub frames in one go. Instead portions of 50 or 100 images can be stacked and finally a stack of stacks is built. This avoids swapping and speeds up the process a bit. Still it will take hours to stack the data on a budget computer. The individual stacks should consist of equal number and images of similar quality.

 

Virtual CN thread:

 

Cheers all, I've got weird stars look like star wars X-wings kinda. Pls help! THX!

 


710x632 pix crop BTW

 

A likely reply:

 

Looks like a tracking problem to me. Get an auto guider!

 

Alas it is wrong. The telescope's image does not show any tracking issues.

 

 

People really try to help beginners. Sometimes it is hard or even impossible to guess what went wrong from a poorly processed image.

 

Next reply:

Did you have a meridian flip during imaging? If your lens suffers from in axis astigmatism the wings are the result of the 180° field rotation during a flip.

 

OP:

??? That's way over my head.

 

For those not familiar with German Equatorial Mounts let me explain that such an instrument can only work in one half of the sky. When the object crosses an imaginary line from the southern horizon through the zenith to the northern horizon, called the meridian, it needs to reverse both axes to continue imaging. That is called the meridian flip. After that, the image is upside down. A tracker or barn door does not need a meridian flip. To explain the star pattern, I stacked the sub frames before and after the flip to individual stacks:

 

  

 

The image is in perfect focus. The comet like flares are called astigmatism. The best lenses reduce this effect to a size that is almost invisible or only visible at the edge of the image. This lens has got plenty of it even in the image center. It clearly indicates that the optical elements are not collimated. (Collimation is the process of adjusting optical elements to each other so that their centers are in a straight line. This process is skipped for cheap lenses simply to keep the price as low as it is.)  Not a big surprise, the fully extended front barrel easily tilts on the slightest touch. As the images are aligned to each other prior to stacking, all images taken after the flip are automatically rotated by 180°. The object is in the same orientation now but the astigmatism is not. Combining both images creates the X-wings. Note that the two stacks show a line pattern in the background which is averaged out in the full stack.

 

The lens also suffers from chromatic aberration. As glass does not refract all colors to the same degree a simple lens does not bring the image to focus in the same place. That creates the blue halos in this image. Some lenses have pink ones. Just buying a mount or a tracker may lead to a great disappointment. The lens that served you so well at daytime is unable to cope with the stars.

 

Let's assume you have a lens that is less than 10 years old and you paid $1300 for it. This must be a lot better than the old cheap Tamron. Here is an example, the Sigma 50-500mm f/4.5-6.3 APO DG OS HSM SLD Ultra. The amount of three letter acronyms and the word "Ultra" may suggest it plays in a different league. Let's give it a try!

 

This was taken on a cheap mount, a stack of 17 x 5 minutes, focal length 500mm f/6.3, scaled down 5 times:

 

Central crop at full resolution:

 

 

Despite the price and the "APO DG OS HSM SLD Ultra", it is not much better. To avoid frustration: even an expensive lens with all sorts of special glass may perform poorly on stars. A telescope for a lower price than this lens will perform much better. The next image was taken with a small quad refractor for about $800. (A refractor is a telescope made of lenses as opposed to mirrors. Quad is short for quadruplet, a construction of 4 lenses that allows for a better corrected image.)

 

 

What is the trick?

Back to my virtual CN thread:

 

Don't use zoom lenses! Primes are better for AP.

 

True. Prime lenses provide just one focal length that cannot be changed. Telescopes are also primes in a way (there are zoom eye pieces, though). A brief look at the price tags: a 200mm prime made by a major brand is available for $600 new. A 135mm with excellent optics but without auto focus is available for $500 new. For lower budgets it becomes very difficult. The best option you have is Ebay and a vintage lens. I got a 200mm f/4.0 Ashai Pentax Super Multi Coated Takumar which was built from the late 1960 to the early 1970s. I got it for $60 plus $20 for an M42 to Canon adapter. Here is an image of the veil nebula that is heavily processed to get rid of the halos, size 5x reduced:

 

 

This is a set of crops from the four corners and the image center at full resolution:

 

 

Finally! This cheap lens is the winner. Cheap now but expensive back then to be precise. There is a 135mm version of the Takumar as well. Prices and condition vary. You may need a lot of patience to find a good one. It does suffer from chromatic aberration but with some effort this can be processed away. Some programs have an automatic CA correction. The manual process is to split the color image to its three color channels, and apply deconvolution to the channel with the largest stars. Deconvolution is the process of shrinking bloated stars back to their original size and shape using measured or estimated data of what the lens is doing to the stars. Free software like RawTherapee offers deconvolution.

 

Conclusion:

Budget AP is possible. When starting from scratch the minimum you need is a tripod, a tracker or a mount, a prime lens and a computer for processing. Most likely you already have a computer, maybe even a camera that is up to the job. If this is the case some luck is needed to find a decent used prime lens. Getting a tracker is not a problem. Skilled tinkerers can build a barn door. The approach needs a lot of patience because the limited exposure time leads to a large amount of sub frames. Results achieved with a zoom lens will be poor even with expensive zooms. More recent 135mm lenses from pre digital age are also available on Ebay. If you want a new lens there is probably nothing under $500. A cheap telescope might be the better option but most telescopes are too heavy for a tracker. You will be limited to the larger and brighter objects. Though not discussed in the article I would like to add that access to a dark place is essential. Under light polluted skies the situation is even worse. Processing the data will be difficult. Processing the clean and deep data from a high-end setup is much easier. It may help to download good data to practice processing.


  • Thomas Pfleger, dswtan, piaras and 13 others like this


9 Comments

Excellent article Elf. I'm sure it'll help a lot of beginners make the right choice when buying a AP rig. I know I made every mistake when I started. Frustrating and expensive. 

 

Regards,

Skyhunter1

    • edl, RichInLesta and alberto76 like this

Very good, thank you!

Photo
Dave Mitsky
May 17 2020 03:55 PM

Well done, Elf!

This article is great. I spent a lot of time researching on what equipment to buy as I am on a budget, and if I had read this when I first started it would've helped me understand exactly why I was buying the equipment I was recommended.

 

Luckily I had done enough research on CN to start off with a pretty good first setup, but the only thing I would've changed; go with a GOTO mount and not just a tracker if the budget allows. A tracker is great, and the Skyguider Pro can get me to 180s, but GOTO abilities are much more attractive than bending over backwards trying to get the target in the frame "just" center.

Yep, Goto can help a lot. For some reason that I do not fully understand virtually everyone on budget photography goes for a tracker although small Goto mounts in the same price range are available. I understand that a tracker is perfect for hiking, getting it to remote places, on airplanes, etc. I do not understand why to use it at home when small mounts are available. Maybe people start out as daylight photographers and search the shops they know. I know old school astronomers take the point a beginner should learn star hopping first and know the sky. Once you learned this you are allowed to take the comfort of a computerized mount. Well, I do not share that view.

    • eros312 likes this
Photo
Andrew Brown
May 28 2020 10:49 AM

Yep, Goto can help a lot. For some reason that I do not fully understand virtually everyone on budget photography goes for a tracker although small Goto mounts in the same price range are available. I understand that a tracker is perfect for hiking, getting it to remote places, on airplanes, etc. I do not understand why to use it at home when small mounts are available. Maybe people start out as daylight photographers and search the shops they know. I know old school astronomers take the point a beginner should learn star hopping first and know the sky. Once you learned this you are allowed to take the comfort of a computerized mount. Well, I do not share that view.

I beg to disagree. winky.gif

 

As can be seen from the amount of "HELP" posts, it is almost always better to learn something from the ground up.

 

Another reason being the amount of gear that gets sold after a season or so.

 

Those are the reasons I always advise binoculars as a first buy. They can be 30, 50mm aperture or 100mm aperture, the commonality being it's up to the user to point them in the right direction, that and their other uses apart from the night sky.  And if the hobby isn't for them the financial losses are less.

Great stuff. Thanks very much.

Great article. You mentioned at the end… It may help to download good data to practice processing.

 

That’s what I want to do. I found data for deep sky objects but not for planetary objects like Moon, Saturn, Jupiter, Venus, Mars. I would rather start my Image Processing Learning and Practicing project with the planetary objects as it seems less demanding than with deep sky objects.

 

I would appreciate it if someone could tell me where I could access free raw data for planetary objects like Moon, Saturn, Jupiter, Venus, Mars.  Many thanks in advance!

Planets raw data are video files and thus pretty large. This may be the reason why you don't find anything.



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics