This post is a bit delayed as I went back and forth on how to frame it (no pun intended), and which topics specifically to cover, because I would like to start showing more of how I create my final planetary images.
My initial thought was to do a grand all-in-one post on the entire process in detail from start to finish. But I ultimately decided that would be a bore to read. Instead, I decided to go with explaining specific units of the process in shorter and hopefully concise explanations.
Today I will focus stacking frames. Planetary (as well as deep sky) imaging requires stacking multiple picture frames, either actual still images or the frames from a video. Because all of these objects are extremely far away, no one single picture can capture enough light to make a full picture. The targets are too small and the number of photons impacted onto any imaging sensor is minuscule. But…if enough single frames are combined in the right way, you can get a semblance recognition of a planet, a star, a nebula, or even another galaxy.
For comparisons, here on Earth, when you shoot a picture, of any type, there are more than enough photons to fill that picture no matter the mode of imaging. Your target is very close to you, measured in feet (meters) or miles (kilometers), it doesn’t matter; deficiency of photons is never going to be a problem.. The same holds true essentially for the Moon as well, on a cosmological scale. In the grand scheme of the Universe, our moon is a stone’s throw from us, ridiculously close, reflecting many photons from a few hundred thousand miles/kilometers away. With the right magnification (a telescope) you can take easy pictures of the Moon with any camera.
On August 23rd I set up my Dobonsian telescope to look at and photograph Jupiter. With a clear sky, I used my normal telescope and camera configuration:
- Telescope: Dobsonian reflector 254mm / 10″ (homemade)
- Camera: Canon EOS Rebel SL3
- Barlow: TeleVue Powermate x5 1.25″
- Filter: Baader Neodymium 1.25″
- Canon T ring and adapter
The computer program post-processing sequence goes PIPP -> Autostakkert -> Registax. Selection of the frames to stack is in the application Autostakkert.
Outside at night, I tried several different ISO and exposure combos. After reviewing all of them, I decided that ISO 1600 and exposure 1/60 seconds was the best that night.
In Autostakkert, you can choose a set number of frames to stack, or a percentage. I always go with a percentage:
You may ask, how does Autostakkert know which frames are best? You have to choose a reference frame, i.e. pick which one you think is the best approximation to what the actual image should be. This is very much art and not science. Here is an example of a completely raw Autostakkert frame for inspection:
Obviously, a lot of picture data is missing from the above image. But it does represent a single source video frame taken from my camera at the telescope.
Once I decided that ISO 1600 and exp 1/60 was the best, I went back to Autostakkert and re-ran the process with different stack percentages. I used 15, 30, 40, 60, and 85 percents.
Here are the finished (non-touched-up) images from Registax:
Typically, for three ~25 second videos, I get about 4500 frames of video. So, for example, the best 30% would be a stacking of 1,350 frames.
My observations on the different percentage stacks are:
- At 15%, the image looks bit grainy, since it is probably still missing some image data to fill the grains in with.
- 30% and 40% are the best, and I have a difficult time deciding which is better. But in the end I decided that 30% looked slightly more clear.
- 60% and 85% are a tad blurry, and that is due to Jupiter’s fast rotation starting to manifest itself. A “best” frame could be at the beginning of the 3-image set or it could be at the very end of the 90 seconds, or anywhere in between. But it’s safe to assume the distribution is roughly normal across the 90 seconds.
So once I had my best ISO, exposure, and percentage of frames stacked, I did some minor post-process editing in PaintShop Pro to (hopefully) sharpen the final picture and (hopefully) reduce noise: