B33 and NGC2024 nebulae

Your neighbour's grass is always g̶r̶e̶e̶n̶e̶r̶ redder

In this web site you will mostly find photos of the Moon (at least so far): in fact, I spent few months between 2020 and 2021 shooting at our satellite, up to making it my specialty a bit (and making me sick of it: in eighteen months I took over 15,000 shots of almost any Moon phase, with a daily and manic frequency).
To say the truth, my curiosity for real astrophotography, that one pointing the camera towards the starry sky and the deep space objects, was born a little earlier, at the end of spring 2020, thanks to two circumstances in relationship only by chance.

The first was to accidentally come across on the internet in a shot of a solar transit of the ISS, the passage of the international space station in front of the Sun, and my wonder in discovering that it was a photo taken with a normal DSLR like the one I own.
The second, the advent of comet Neowise in our sky, the protagonist of summer 2020, caught in millions of photographs relaunched by all social networks, taken even with simple mobile phone cameras.
I shot (badly) at both of the above events, but I will write about them on another occasion. Here I want to tell about the latest work that I have just completed, which represents a bit the synthesis of my path for almost two years now in the field of astrophotography.

The opening image above in this article represents one of the most popular and photographed pairs of nebulae in our sky: NGC 2024, known as the Flame Nebula, and B33, better known as the Horsehead Nebula due to its characteristic shape (NGC and B are used as a label to classify the so-called "deep sky objects", that is codes used in astronomy identifying almost everything that is observable in our universe).
To obtain this image I spent ten days of work (and seven nights) in taking and processing a huge number of digital photos. The extraordinary thing about this image is that it portrays something staying more than a thousand light years away from us and, above all, that it's been taken without using a telescope, simply with a normal DSLR and a little extra equipment, not strictly indispensable, but which certainly helps.
Unbelievable, isn't it?

Barnard 33, the Horse Head, is one of the first DSOs that I was passionate about. It is an image frequently recurring when talking about astrophotography, partly due to its characteristic and suggestive shape, partly because it is one of the easiest objects to locate in our skies without having to rely to ad hoc tools (although, unlike other nebulae and galaxies, it is almost impossible to find it without a good camera), partly because the region of the sky where it is located, the Orion constellation, more than being extremely popular also among people who normally does not deal with stars and celestial bodies, it's one of the richest of objects that can be easily taken with an ordinary camera.
Anyway Horse Head is actually not so simple to be caught: although you can easily drive yourself among the stars to find the area where it's located, to shoot at it successfully it's definitively a different matter, especially where the sky is mostly light polluted.

Map of Orionis

How to locate B33 and NGC 2024 nebulae in Orion

There are objects in our sky much easier to be taken with a camera by a beginner, sometimes also visible to the naked eye (provided the sky is clear enough): for example, still exploring the Orion region, the famous M42, also known as the Great Orion Nebula, so bright that it be clearly distinguishable on clear nights even without binoculars; or even the Andromeda Galaxy, which in the absence of light pollution appears as a very bright orange star surrounded by a blurred glow - provided you know where to look for it and how to identify it.
These objects can be "photographed" with a single long exposed shot taken with any camera and a regular lens. I write "photographed" in quotation marks to mean that, in this case, the result will certainly not be comparable to what you see in this post, but anyway distinguishable in the frame.
M42 can be taken even with a mobile phone, provided the sky is perfectly clear, using a tripod and being satisfied by a photo showing only a very bright spot rather large in the middle of other stars.

B33 and the nearby NGC 2024 are invisible to the naked eye, much smaller compared to the size of the previous objects, and above all much less bright: even using a good camera and a long enough exposure it is not possible to identify them within a single shot, except for exceptionally clear skies where they can appear as a very faint and faded grayish spot, barely distinguishable from the darkness of the surrounding deep space.
Despite this, they are one of the most coveted targets due to their easy localization combined with the possibility of taking them without using telescopes and the indisputable fascination they exert on the imagination: if you want to amaze friends by photographing suggestive colored nebulae in the deep sky, Horse Head and Flame are the perfect subject needing only a minimum of equipment, skills and experience, for them being not possible to shoot at by simply getting out on your terrace and taking a pic with your mobile phone.

Therefore, after my first attempts with easier and more traditional objects (M42 and Andromeda, of course), having made a minimum experience of basic notions and adopted some technical measures to update my photographic equipment, in the autumn of 2021 I too started hunting for Barnard 33, which quickly became my first real challenge to face with advanced amateur astrophotography.

To get the image of this post it took me six months of in-depth study of the technique and software to be used (always the infinite resources of the internet be blessed), unsuccessful attempts, nights spent in the cold fighting with the photographic equipment to frame the nebulae, extreme frustration, time irremediably wasted in wrong procedures, few quick wins achieved step by step and progressive adjustments of my equipment with significant hacks.

B33 acquisition progress

miei progressi nei tentativi di fotografare la Testa di Cavallo

One of the characteristics of astrophotography is that it often takes hours, or even days, to understand if you are proceeding the right way and being able to see the result of your work. You often move forward blindly through the image processing steps, mostly based on your previous experiences and/or on what you read around on the web.

The key point in the process is that you're shooting at things that you actually "don't see". So, unless you own an expensive equipment able to do the dirty work for you by automatically pointing the camera at your target, you will need lot of patience, study, and luck to center your target and getting what you want.
In the case of Barnard 33, pointing the camera where the nebula is presumably located is quite easy; being sure and able to keep the framing for hundreds of shots, especially during consecutive evenings, is another story.

The first time I was able to "see" the B33 picture I published here it was after a week of image processing work, the last step of which took over twenty-four hours of calculation depending on some parameters that I set up based only on my previous few attempts with other DSOs.
The result I got, at that point, turned out to be a disaster. I had a moment of deep despair and quite the temptation to throw everything away.
The problem was that I had no idea whether that disaster was due to any of the parameters set for that last procedure, which would have meant rolling the image processing twenty-four hours back, or - for instance - depending on the mode I shot the photos, which in that case would have meant throwing away many days (and nights) of work.
It took me a whole day just to get an idea of ​​where I went wrong and decide from which point of the procedure to start over.

Among many problems to face when shooting at objects such as nebulae and galaxies, the first you cope with is the need to store and deal with huge quantities of data due to the way DSOs are long exposed by accumulating hundreds of shots, with the aim to capture as much as possible the very faint light they emits.
In the simplest terms, the goal is achieved by "summing" a large number of photos until an exposure time of several hours is reached. The longer the overall exposure time, the better the quality of the final result.
There is not a time limit, nor an ideal one: you can get good results with one hour of total exposure "integration time" if you're shooting at very bright objects with clear sky, or it could take fifteen, twenty, forty hours of total exposure when targeting objects whose brightness is very faint, or while shooting under heavily light polluted skies.

The image in this post is obtained from a set of shots whose overall exposure time is over twelve hours. This can basically be achieved in two ways: by lengthening the exposure times of each photo trying to reduce the overall number of shots to be taken, or increasing the total number of shots. Usually the best compromise is sought between the two solutions according to many variables that essentially depend on the equipment you have (photographic and computational for image processing) and on the location from where the photos are taken.
You must remind that the more shots you're going to use, the more time and resources will be needed for their storage and processing. When dealing with astrophotography it is a while to get to 1Tb of data to be managed.
This is why it is important to do everything possible to lengthen the exposure times and reduce the number of images to be processed, without sacrificing the quality of the final result.

Another key factor you must consider when shooting with long time exposures at objects in the sky, be they the Moon, planets, or stars, is that their apparent motion must be taken into account: the longer the focal length of the lens you are using, the quicker the objects move across your frame and lengthening the exposure times even by fractions of a second result in blurred photos

To get photos like the one in this post and to extend the exposure times as much as possible, you basically need "to chase" the stars in their motion through the sky, that means making the camera to follow the apparent motion of the starry sky while keeping the framing aligned at your target. In other words, the stars will not have to move around the frame as they are being shot.
In case you do not have an adequate stuff to chase the stars, you can find many resources on the web helping you to calculate the maximum possible exposure time according to your camera and the lens used so that stars remain pinpointed without leaving trails. Generally, you will need to deal with exposure time around 1s or a little more.
This is the main reason why in astrophotography "astro trackers" are used, that is particular rotating mounts on which camera is attached making it to follow the stars motion so that stars remain pinpointed even with prolonged exposures.

Astro-trackers are very sophisticated tools, but they are not necessarily very accurate and, as usually happens, the more you want to save money, the less the tracker will accomplish its task perfectly (which does not mean it will not simplify your life in a very significant way).
A basic astro-tracker, like the one I use, can help you in increasing the exposure times up to a few minutes before the stars leave light trails, at least with short focal lengths and wide-angle lenses. Basically they are absolutely perfect tools to get pictures of the Milky Way in the mountains, even with a single click.
By increasing the focal length of the lens (and the weight of the equipment mounted on the star tracker) the engine tends to struggle in following the motion of the stars and the potential exposure times are significantly reduced, even if they always remain quite over what you can do with your camera mounted on a normal tripod.
As an example, to get the Horse Head image I published in this post I used a 600mm zoom: with my astro-tracker I can expose a single photo up to 20" before loosing the star alignment and the stars leave trails; without a star tracker 1" of exposure time would be enough for the trails to come out.

Another intrinsic feature of the astro-tracker is that regardless of the exposure time set for the photos you will not need to adjust the framing all along the photographic session, since it remains centered on your target for several hours and hundreds of shots. Indeed, using a normal tripod, in addition to struggle with shorter exposure times you will have to realign the framing every few minutes: a nightmare if you are shooting at easy and very bright objects with wide focal lengths; almost impossible with long focal lengths and deep space objects.

Once you have found the right compromise among exposure time, ISO value (too high means large amount of light pollution and background noise affecting your shots, too low means not enough light caught), aperture value, focusing, weight balancing and polar alignment of the tracker, you can start collecting data, i.e. taking hundreds, or even thousands of shots keeping your target framed, with the aim of getting an exposure time as long as possible, potentially of several hours, in order to store as much light, and therefore information, as possible.

The more you can extend the exposure times without light trails appearing in the photos, the fewer photos are needed; the shorter the exposure times, the more photos you need.
The longer the overall exposure time, the better the quality of the final result.
The more photos you accumulate, the more processing times and storing resources you will need (disk space, CPU, etc.).
In short: astrophotography is a continuous search for the best combination of skills and resources in the photographic field, and skills and resources in the computational field.

On the other hand, a definitive result is never achieved: it is always possible to get further data to be added to the old one, even after months, therefore increasing the total exposure time and improving the quality of the final result.
As far as I'm concerned, however, although I've undoubtedly improved a lot, the photos of B33 taken by other astrophotographers I follow on Instagram and Reddit are always (much) better than mine. It is still not clear to me whether this is an incentive, or rather a source of extreme frustration, considering all the time spent so far to be able to obtain these results.

Some technical data. The image in this post has been obtained using a Canon 90D equipped with an Optolong L-Pro narrowband filter to reduce light pollution. I mounted a Sigma Contemporary 150-600mm lens at its maximum focal length.
The actual framing referred to the sky and the Orion constellation is represented by the red rectangle drawn on the previous image.

As a star tracker I used a Skywatcher StarAdventurer 2i hacked by adding a 2.5kg counterweight more in addition to the standard 1kg, perpendicularly mounted with reference to the standard bar, in order to balance the zoom weight. Overall, the equipment mounted on the star tracker weighed about 8kg, definitely at the limits of the StarAdventurer's capabilities.

The final image was obtained by registering and stacking 2166 shots selected from over 4000, taken at ISO1250, f/6.3, with an exposure time of 20s, for a total of more than twelve hours of integration time.
The photos have been taken over seven nights (in the current season Orion already appears high in the sky immediately after sunset and sets in the West around midnight; considering the light pollution of the area where I live, it can be photographed at most in the interval 7pm-11pm approximately).

In addition to the "light frames", that is the actual photographs, I also acquired 97 flat field frames and 133 bias frames in a single session during the first night, and 311 dark frames distributed all along the seven nights, in order to calibrate the photographs and improve the signal to noise ratio.

The total data volume, including all images due to intermediate processing steps, is greater than 1.2Tb.

The registering and stacking processes were done on Siril on a quad-core 2013 Intel iMac connected to a QNap NAS, and took 36 hours approx.
The post processing steps were performed with PixInsight, Photoshop and Topaz AI Denoise on a 2021 8+8 core MacBook Pro M1.

B33 region

My equipment

Leave a comment