Thanks, guys! 🙂 Really glad you like it. I have more on the way, too! Tonight ended up being clear again (unexpected, but a happy surprise.) I’m imaging M78, also in Orion, right now. Much smaller, much dimmer object. Will be more challenging, but I think it’s a very beautiful reflection nebula. So stay tuned!
Thanks! I’d be happy to share. I will probably write some articles on the subject when I am more skilled, bit in the mean time, I’ll give you the basics here.
First off, I used this setup: https://jonrista.com/2014/02/13/atlas-eq-mount-600mm-telescope/. To get an image like this, the most important piece of equipment is having a tracking mount. If you have the ability to track the stars across the sky, then you can actually to a LOT with your standard camera gear…a DSLR and a 100mm or longer lens will get you some great shots. You can find things like the iOptron SkyTracker, which can hold maybe 7 or 8 pounds, and is able to track accurately enough for 100-200mm lenses. You could also get what I did, something like the Orion Atlas EQ-G, which will hold much heavier gear.
Once you have a tracking mount, the first part of the process is to get the images. This particular image is a stack of 30 180 second individual exposures. That is a total of 5400 seconds of exposure, or 90 minutes (hour and a half.) The reason you do that is because astro imaging, unlike normal photography, usually results in very dim photos. The dynamic range of an image like this is HUGE, because the stars are exceptionally bright, while the nebula details are exceptionally dim. That means most of the really interesting stuff, the nebulas, is in the bottom few bits of shadow detail…stuff that normally looks black in your average photo. When you stack a bunch of individual frames and use a median averaging algorithm to blend them together, that removes a whole lot of the noise.
In addition to simply taking a bunch of “light frames”, you also need to take dark frames, bias frames, and flat frames. I’ll go into this stuff in more detail once I get around to writing about this stuff in my knowledge center. Suffice it to say, these allow you to remove the noise added by the camera itself, allowing the background sky (the space between the stars) to be even darker and blacker, without banding or fixed patter noise issues.
Once the noise is gone, then you have the ability to “stretch” the image. Stretching is a process by which you lift up that ultra dim shadows where that nebula detail starts, and make it bright enough to see. There is a lot of information on the internet about stretching and processing astro images. The simplest form usually involves the levels and curves tools in Adobe Photoshop, and a little bit of experimentation will usually give you a general idea of how to stretch enough that you start getting some good detail out of the depths of the shadows. After you’ve stretched the image, then you usually have to apply a bit of noise reduction and color correction, and maybe one final post-NR stretch, to really bring out all the detail in accurate color.
Regarding this image of Flame and Horse Head nebulas. I probably had about 1/3 to 1/4 as many light frames as I needed. I stacked 30, I really needed 90-120. If I manage to get another clear night before Orion sets for good, I’ll try to get some more frames, and maybe I’ll be able to re-stack this image and get cleaner results. The more you stack, the less noisy the image will be, and the easier it is to lift up deeper and deeper shadow detail. (You can’t really see it in my shots here, but the entire region between flame and horse head, surrounding IC2023, is all dark nebula that has this look like a rippling satin sheet…some day I hope to expose my astro images well enough that I can lift up even that kind of detail. That’s a bit beyond my skill at the moment, however…and I get the feeling, with the weather as it is, I might not get the chance until Fall 2014.)
Thanks Jon, that’s interesting stuff. I’ve played around with deep sky stacker a little, but always ended up with less than ‘stellar’ results. I still don’t quite understand what bias and flat frames are. Also, without the tracking mount, I’ve had to use shorter exposures to avoid star trails and then end up with a much smaller image after alignment.
Without a tracking mount, your pretty limited. Shorter lenses help a lot for untracked, so very wide field or whole sky is the way to go. If you do whole sky, you don’t necessarily need to stack, although stacking can still help. You can expose on a FF camera for up to around 30 seconds, on an APS-C crop camera to maybe 20 seconds at the most, with an ultra wide angle lens like a 10mm – 16mm. Your maximum exposure time drops from there as focal length increases. The best way to figure out how long you can expose is the 500 rule: 500/focalLength will give you the absolute maximum time you can expose without trails. So, at 100mm, you can expose for a mere 5 seconds (and in practice, 5 seconds is still a bit too long, it often results in trails.) If you expose for 3-4 seconds at 100mm at ISO 800 or 1600, and stack about 150 exposures (450-600 seconds, or 7.5-10 minutes), you’ll finally start getting some better results.
The trick, however, is to only take a few exposures then re-center. If you do not re-center, the movement of the sky will take the region your imaging out of the frame, and when you stack, you might only get a sliver of clean results. At 100mm, which is enough to image say Orion’s Belt and Sword, or most of Cassiopeia, you can take maybe three to four exposures before you need to re-frame. The best way to frame is to identify a region of the sky that kind of fits within the first box of your viewfinders smallest grid display. That gives you a margin around the area of sky you want to keep, allowing some movement that won’t become a problem. Frame the region of sky such that the area you want to keep starts fairly close to the leftmost edge of the frame, but not too close to it that it’s actually touching. Then watch, and see how that bit of sky moves across your frame. The stars will drift from left (east) to right (west), but will also drift either up or down. If they drift up in the frame, recenter such that the stars are as far to the left edge and as low down as you can get, without cropping out any part of the region you want to capture (i.e. if you want orion’s belt, make sure his entire belt is always within the frame.) If the stars drift down in the frame, recenter such that the stars are as far to the left and as high up as you can get. Once you have the orientation right, let the stars drift again, and calculate how long it takes for any part of the region of sky you want to image to finally move out of the frame on the opposite side. That time is your sub-sequence time.
You can set up your camera with a remote shutter release and intervalometer that can automatically take pictures to take N*(s+d) pictures, where that total is less than your sub-sequence time. If it takes 3 minutes for the sky to start drifting out of the frame, and your exposing for 4 seconds, and your delay between frames is 1 second (most intervalometers require some minimum delay between shots), then you can take at most 36 subframes before you have to reframe. Set up your intervalometer to take 35 4-second shots with a 1-second delay between frames, and let ‘er rip. Now, it’s just a matter of getting the necessary number of frames to stack effectively. At 35 frames a sub-sequence, if you want 150 frames, you need five sequences (that will give you 175 frames). Once you stack, you should find that you have FAR less noise. With less noise comes the ability to stretch those dim nebulosity details.
There are a few caveats with untracked sequencing like this, or even sequencing with an Alt/Az tracker. You get field rotation. If you think about the orientation of the constellation Orion when it rises in the east, he is “standing” on the south, with his head pointed north. As the night moves on, Orion remains in this orientation relative to north and south as he moves across the sky, however his orientation to any rectilinear field of view changes. If you point your camera, level to the ground, at Orion in the east, his head will point towards the left edge of the frame. When directly overhead, again if you point your camera, level to the ground, strait at orion his head will point towards the top edge of the frame. By the time he is over the western horizon, when frames in-camera his head will point towards the right edge of the frame. This is field rotation. It’s the reason we have equatorial tracking mounts…as the way they track conforms to the celestial coordinate system, so the orientation of the scope (and therefor the camera sensor frame) is continually reoriented to ensure that as you track something like Orion across the sky, he is always oriented exactly the same in the frame…no field rotation.
Field rotation complicates the alignment process, however a tool like DSS will align for you. The result of field rotation is that things don’t always line up exactly right, so stars get a little bigger, they are often not exactly round and instead they are misshapen and amorphous. You lose a little bit of detail in nebula and galaxies. It is possible to get some relatively cheap tracking devices for cameras and smaller lenses without having to spend the big bucks on a big mount. Something like the iOptron SkyTracker can hold up to around 7 or 8 pounds, can be polar aligned, and track equatorially. They cost a few hundred bucks, but they can completely change the game for ultra wide field astrophotographers, allowing you to expose for up to a minute or two (assuming you polar align it properly and accurately), without field rotation.
Nature Photography 
Thanks, guys! 🙂 Really glad you like it. I have more on the way, too! Tonight ended up being clear again (unexpected, but a happy surprise.) I’m imaging M78, also in Orion, right now. Much smaller, much dimmer object. Will be more challenging, but I think it’s a very beautiful reflection nebula. So stay tuned!
Truly outstanding! Well done Jon!
Jon, that is amazing, care to share the process that it took to create this?
Thanks! I’d be happy to share. I will probably write some articles on the subject when I am more skilled, bit in the mean time, I’ll give you the basics here.
First off, I used this setup: https://jonrista.com/2014/02/13/atlas-eq-mount-600mm-telescope/. To get an image like this, the most important piece of equipment is having a tracking mount. If you have the ability to track the stars across the sky, then you can actually to a LOT with your standard camera gear…a DSLR and a 100mm or longer lens will get you some great shots. You can find things like the iOptron SkyTracker, which can hold maybe 7 or 8 pounds, and is able to track accurately enough for 100-200mm lenses. You could also get what I did, something like the Orion Atlas EQ-G, which will hold much heavier gear.
Once you have a tracking mount, the first part of the process is to get the images. This particular image is a stack of 30 180 second individual exposures. That is a total of 5400 seconds of exposure, or 90 minutes (hour and a half.) The reason you do that is because astro imaging, unlike normal photography, usually results in very dim photos. The dynamic range of an image like this is HUGE, because the stars are exceptionally bright, while the nebula details are exceptionally dim. That means most of the really interesting stuff, the nebulas, is in the bottom few bits of shadow detail…stuff that normally looks black in your average photo. When you stack a bunch of individual frames and use a median averaging algorithm to blend them together, that removes a whole lot of the noise.
In addition to simply taking a bunch of “light frames”, you also need to take dark frames, bias frames, and flat frames. I’ll go into this stuff in more detail once I get around to writing about this stuff in my knowledge center. Suffice it to say, these allow you to remove the noise added by the camera itself, allowing the background sky (the space between the stars) to be even darker and blacker, without banding or fixed patter noise issues.
Once the noise is gone, then you have the ability to “stretch” the image. Stretching is a process by which you lift up that ultra dim shadows where that nebula detail starts, and make it bright enough to see. There is a lot of information on the internet about stretching and processing astro images. The simplest form usually involves the levels and curves tools in Adobe Photoshop, and a little bit of experimentation will usually give you a general idea of how to stretch enough that you start getting some good detail out of the depths of the shadows. After you’ve stretched the image, then you usually have to apply a bit of noise reduction and color correction, and maybe one final post-NR stretch, to really bring out all the detail in accurate color.
Regarding this image of Flame and Horse Head nebulas. I probably had about 1/3 to 1/4 as many light frames as I needed. I stacked 30, I really needed 90-120. If I manage to get another clear night before Orion sets for good, I’ll try to get some more frames, and maybe I’ll be able to re-stack this image and get cleaner results. The more you stack, the less noisy the image will be, and the easier it is to lift up deeper and deeper shadow detail. (You can’t really see it in my shots here, but the entire region between flame and horse head, surrounding IC2023, is all dark nebula that has this look like a rippling satin sheet…some day I hope to expose my astro images well enough that I can lift up even that kind of detail. That’s a bit beyond my skill at the moment, however…and I get the feeling, with the weather as it is, I might not get the chance until Fall 2014.)
Thanks Jon, that’s interesting stuff. I’ve played around with deep sky stacker a little, but always ended up with less than ‘stellar’ results. I still don’t quite understand what bias and flat frames are. Also, without the tracking mount, I’ve had to use shorter exposures to avoid star trails and then end up with a much smaller image after alignment.
Without a tracking mount, your pretty limited. Shorter lenses help a lot for untracked, so very wide field or whole sky is the way to go. If you do whole sky, you don’t necessarily need to stack, although stacking can still help. You can expose on a FF camera for up to around 30 seconds, on an APS-C crop camera to maybe 20 seconds at the most, with an ultra wide angle lens like a 10mm – 16mm. Your maximum exposure time drops from there as focal length increases. The best way to figure out how long you can expose is the 500 rule: 500/focalLength will give you the absolute maximum time you can expose without trails. So, at 100mm, you can expose for a mere 5 seconds (and in practice, 5 seconds is still a bit too long, it often results in trails.) If you expose for 3-4 seconds at 100mm at ISO 800 or 1600, and stack about 150 exposures (450-600 seconds, or 7.5-10 minutes), you’ll finally start getting some better results.
The trick, however, is to only take a few exposures then re-center. If you do not re-center, the movement of the sky will take the region your imaging out of the frame, and when you stack, you might only get a sliver of clean results. At 100mm, which is enough to image say Orion’s Belt and Sword, or most of Cassiopeia, you can take maybe three to four exposures before you need to re-frame. The best way to frame is to identify a region of the sky that kind of fits within the first box of your viewfinders smallest grid display. That gives you a margin around the area of sky you want to keep, allowing some movement that won’t become a problem. Frame the region of sky such that the area you want to keep starts fairly close to the leftmost edge of the frame, but not too close to it that it’s actually touching. Then watch, and see how that bit of sky moves across your frame. The stars will drift from left (east) to right (west), but will also drift either up or down. If they drift up in the frame, recenter such that the stars are as far to the left edge and as low down as you can get, without cropping out any part of the region you want to capture (i.e. if you want orion’s belt, make sure his entire belt is always within the frame.) If the stars drift down in the frame, recenter such that the stars are as far to the left and as high up as you can get. Once you have the orientation right, let the stars drift again, and calculate how long it takes for any part of the region of sky you want to image to finally move out of the frame on the opposite side. That time is your sub-sequence time.
You can set up your camera with a remote shutter release and intervalometer that can automatically take pictures to take N*(s+d) pictures, where that total is less than your sub-sequence time. If it takes 3 minutes for the sky to start drifting out of the frame, and your exposing for 4 seconds, and your delay between frames is 1 second (most intervalometers require some minimum delay between shots), then you can take at most 36 subframes before you have to reframe. Set up your intervalometer to take 35 4-second shots with a 1-second delay between frames, and let ‘er rip. Now, it’s just a matter of getting the necessary number of frames to stack effectively. At 35 frames a sub-sequence, if you want 150 frames, you need five sequences (that will give you 175 frames). Once you stack, you should find that you have FAR less noise. With less noise comes the ability to stretch those dim nebulosity details.
There are a few caveats with untracked sequencing like this, or even sequencing with an Alt/Az tracker. You get field rotation. If you think about the orientation of the constellation Orion when it rises in the east, he is “standing” on the south, with his head pointed north. As the night moves on, Orion remains in this orientation relative to north and south as he moves across the sky, however his orientation to any rectilinear field of view changes. If you point your camera, level to the ground, at Orion in the east, his head will point towards the left edge of the frame. When directly overhead, again if you point your camera, level to the ground, strait at orion his head will point towards the top edge of the frame. By the time he is over the western horizon, when frames in-camera his head will point towards the right edge of the frame. This is field rotation. It’s the reason we have equatorial tracking mounts…as the way they track conforms to the celestial coordinate system, so the orientation of the scope (and therefor the camera sensor frame) is continually reoriented to ensure that as you track something like Orion across the sky, he is always oriented exactly the same in the frame…no field rotation.
Field rotation complicates the alignment process, however a tool like DSS will align for you. The result of field rotation is that things don’t always line up exactly right, so stars get a little bigger, they are often not exactly round and instead they are misshapen and amorphous. You lose a little bit of detail in nebula and galaxies. It is possible to get some relatively cheap tracking devices for cameras and smaller lenses without having to spend the big bucks on a big mount. Something like the iOptron SkyTracker can hold up to around 7 or 8 pounds, can be polar aligned, and track equatorially. They cost a few hundred bucks, but they can completely change the game for ultra wide field astrophotographers, allowing you to expose for up to a minute or two (assuming you polar align it properly and accurately), without field rotation.
Jon, these are fantastic photos. You are very talented!
Thanks, Sue! Pleased to hear you say that.