Astro Photography

Both the image of the Orion Nebular and the one from the Pleiades were captured from my home location under moderate conditions (Bortle 5, Light Pollution Map Karben) including light pollution and a typical winter dust. These pictures were never intended for publishing and were taken as a first light with a new telescope to check focus and guiding quality of the equatorial mount.

But this exact image is a nice example to demonstrate, what astro photography is all about. Regardless which type of camera was used to capture this picture, to obtain a colorful and detailed result some massive tuning on the gradation curve will be required. Only then those fine structures and gradients will reveal. After application of a quick "astro"-curve we may get a result like this:

Only that extreme curve shown at the left was applied to the original image, any we already recognize good colors and contrast. The curve actually starts with the light pollution (the left peak), which already throws away about half of the brightness values. Then the steep slope stretches about 10% of all brightness values across the full range (from black to white). For this reasont this step is called "stretching".

Everyone familiar using these curves is aware of the fact, that such massive adjustments reveal image artefacts like sensor noise. On the right you'll see an enlarged crop from the center for better visualization.

Sad point is, that this noise can not be controlled using exposure time only. In this example a longer exposure time would increase the base brightness from light pollution in first place. Then the already too bright core would get even more overexposed. In the end this will reduce the usable brightness range hence decreasing image quality. On the other hand a reduced exposure time would make sensor noise more dominant.

The relevant keyword in astro photography is Signal-to-Noise-Ratio (SNR), effectively the distance between the signal of interest and the omnipresent noise-floor. We simply need to maximize that signal to noise ratio. For a single exposure the SNR is basically defined by the bit resolution of the sensor (quantization noise), the number of brightness levels which can be differentiated. For JPG files this would be 8 Bits (per color), so we may differentiate only 256 brightness levels. Raw image, in contrast, provide at least 4096 (12 Bit), most of the times 16384 (14 Bit) or even 65536 (16 Bit) brightness levels (per pixel and color channel), which result in a less noisy image after applying such a curve. But the source of this image originates from a 14 Bit RAW file of the D750 (specialized astro cameras not necessarily have a higher bit depth) and still does not look nice.

Now data processing takes over and we need many more images from the same object. The individual pixels of all images will be integrated and the resulting color value evaluated. Since sensore noise occurs more randomly it is reduced in this step the more images are processed. This not only compensates the limited bit depth of the camera sensor, but removes other noise like satellite trails as well.

After stacking 28 of these images along with flats and darks and some later corrections, the result may look like this (D750, 580mm, f/5.8, ISO1600, 28x 120s):

From a mathematical standpoint a duplication of the available data quantity halves the quantization noise. The tiny series from the region horsehead nebula illustrates this pretty well.

As an experiment I took four exposures of 300 seconds each (ZWO ASI2600MC Pro) during full moon right into the light pollution of Frankfurt. The series of 1, 2, 3 and four integrated images was then stretched in Photoshop. With this little quanity of images the quantization not only the noise decreases rather quick, the finer structures get more pronounced as well. For a similar significant further reduction, we'd need 8 images, then 16, 32, 64 and so on. 60 images of 300 seconds equals a total exposure time of 5 hours. This is where these long exposure times originate from in astro photography.

I remember being quite excited to manage capturing an image at all of something that far away from us. But the result was way off from what I expected to get so focus was lost on this topic.

Nowadays, some 15 years later, that image probably still arouse pity. When searching the internet for astro images you will found plenty of great shots which sometimes were taken under much worse conditions. Specialized software tools for astro photography evolved a lot since. They seem to do a much better job than in those past times and, of course, I've learned a lot as well.

I intentionally presented that original image here to clarify, that taking some exposures is not the end of the story. In fact it is the first step within the workflow. A nice image needs some attention in post processing and editing. So I decided to take the original images (no correction frames) and push them through a modern workflow. So this is the result (identical framing):

Of course this still is not an exceptional image, but clearly demonstrates the capabilities of modern imaging tools, even if there are only compressed JPEG files without correction frames available.

After realizing this, some curiosity came up. So I picked one of my telescopes, a somewhat reasonable 6" reflector telescope with a focal length of 750mm and f/5 aperture. The only goal was to determine, if there at least is a bit of a chance to image some deep sky objects from my home. If this could prove at least some success, there was a chance to revive that abandoned hobby.

The obtained result was never intended to be presented anywhere, since I was totally aware about the shortcomings in this setup. But to provide you a better idea of what to expect, or better, not to expect from such a standard telescope, which still is great for visual observations, I'll show it anyway (D750, 28x 120sec, ISO3200):

On first sight this does not look that bad, except for some dust on the mirrors. But after that initial impression some severe focus issues are pretty obvious. One issue is the mechanical stability of the eyepiece extension. As soon as something heavy is attached, like a DSLR, it bends a little so the focal plane is tilted slightly. As a result we get a focal gradient across the image.

While this probably could be fixed mechanically, a loss of focus towards the corners originates from the simple construction of this reflector telescope. An uncorrected "Newtonian" telescope creates a spherical focus at the eyepiece, which is ok for visual astronomy, but highly useless for astro photography, which requires a focus plane. Both issues can not be corrected afterwards.

Depending on the type of telescope different countermeasures exist to generate a proper focal plane for astro photography. In general this is realized with some more or less complex lens systems. And, like any kind of specialized optical system, these tend not to be cheap. If you intend to purchase a new telescope for astro photography and try to save some bucks here, you'll be "rewarded" with images like the one above.

It took me a while to decide for a new telescope with regards to optical quality, targets of interest and budget. In the end I now own a, so called, apochromatic refractor with a focal length of 580mm and an aperture of f/5.8 (hence a diameter of 100mm). Images taken with this telescope are on quite a different level and razor sharp up to the corners. Here is Andromeda again, taken from the same spot as the previous image (D750, 32x 120sek, ISO640):

Specialized astro cameras exist in two main variants, color and monochrome. I'll explain the major aspects on a separate page: Color or Monochrome?

With two years of experience, a different telescope and a monochrome camera I've imaged that same region again in late sommer of 2023, the new image can be found here: Around the Pelican Nebula

My current setup looks like this:

• Teleskop: TS-Optics TSQ-100ED 580mm focal length, f/5.8, apochromatic refractor
• Teleskop: Sharpstar 13028HNT "Hypergraph 130", 364mm focal length f/2.8, hyperbolic Newton
• Mount: Skywatcher HEQ-5 Pro
• Polefinder: ALCCD QHY PoleMaster
• Guiding-Camera: ZWO ASI120mm Mini
• Main-Kamera: ZWO ASI2600MM Pro (26MPixel, APS-C) with ZWO Filterwheel
• Filters: LRGB: Antlia Pro, narrowband: ZWO 7nm
• Alternate Kamera: ZWO ASI2600MC Pro (26MPixel, APS-C)
• Fokuser: ZWO EAF Motorfocus
• Control computer: ZWO ASIair Plus
• Power supply: Jackery Explorer 500
• Stacking-Software: AstroPixelProcessor
• Editing: PixInsight, Photoshop CC, Noise/Blur/StarXTerminator, Topaz AI

The dominant amount of ZWO components was set by my decision to go for a ZWO ASIair, which (at the moment) only supports components from the own eco-system (except for mounts). But there are high quality alternatives without such restrictions like the open source software N.I.N.A..

With this setup I am able to capture many medium sized deep sky objects with a very decend resolution. But since the camera has an APS-C sensor, the setup effectively has a focal length of 870mm, so larger objects like the Andromeda galaxy do no longer fit into a single frame. For such objects I still may use the DSLR. On the other hand that focal length is still way to short for distant galaxies or planets.

To determine your favorite focal length you may use tools like Telescopius.com or Stellarium. These provide the function to enter both focal length of a telescope and a camera (sensor size) and to place the resulting frame rectangle into the virtual sky.

This final image is a detail of the constellation cygnus within the milky way. With the bright star Sadr bottom right, there is the Crescent nebula (NGC6888) on the left and the star cluster M29 right on top. All surrounded by a massive amount of stars from the milky way. Taken with the D750 at ISO640, integrated from a total of 70x 118 seconds subframes, a total of about two hours exposure time: