Riding The SNR Curve
Posted: Sat Apr 03, 2021 4:51 pm
I was fortunate to have had two clear nights in a row, a perfect opportunity for experimentation. I have a mono camera and filters. I like the flexibility, but it also presents some challenges. My biggest challenge is my refractor. It does not focus blue very well. I've learned that conventional LRGB imaging results in bloated blue stars, but I discovered a remedy. I simply eliminate blue by using a Wratten #12 filter (minus blue) for luminance, and standard red and green filters. So, why do my images show blue? (See attachment.) It is the way that I mix the channels in Astro Pixel Processor (APP).
At first, I had trouble processing these images of M51. I was following the same technique I use for star clusters, but I was not getting the result I expected. The image had a severe red bias. In a topic that I started a few days ago, entitled "Exposure's Effect on Color Quality," I speculated that the exposures I used were not long enough. I was wrong. The problem was with APP. It turns out that I need to run Star Color Calibration on galaxies when I use the Wratten #12, red, and green filters. Now that I have that behind me, I can move on with my analysis.
Looking at the attachment, your first reaction might be "I like the left-hand image best." I don't disagree with you. I like the HII regions in the arms (the red highlights.) The right-hand image has some red, but it is muted. The reason for that is due to the order of filters. East of the meridian, I start with red, then green, and finally Wratten #12. Red light wavelengths pass through the atmosphere more easily than green and blue at a given altitude above the horizon due to Atmospheric Extinction. The unfortunate consequence of starting with red is that the atmosphere is still unsteady after sunset. So, in the right-hand image, the atmosphere was turbulent when I captured reds. However, on the night before, the left-hand image, the atmosphere was steadier, and therefore red light was stronger.
If we overlook the loss of red highlights in the right-hand image, we can see that the galactic core and arms are brighter. How can this be? I spent 151 minutes on the left-hand image, but I got a worse result. The right-hand image took only 110 minutes, but it is so much brighter.
Oh no! There goes Brian again with his 300-second exposures. Let me say, that 300 seconds is completely arbitrary. I just wanted to make it significantly long enough so that the difference was obvious. Furthermore, I have a paltry 71mm aperture. A lot of you guys and gals have 8 inches and greater. You are scooping up far more photons than I. I don't doubt that you are saying to yourself, "Brian, that's a pretty cr**py image for 2 hours of effort." Yes, that's what I get for having a small aperture. Believe me, if I could start all over again, I'd go straight for an 8-incher.
The point that I want to make is that you can get a brighter image in less time if you increase exposure. And there is the proof. Now of course there could be many reasons why you don't want to go longer. Among those are: severe light pollution, poor guiding, dodging clouds, and aircraft. Those are all valid reasons. Thankfully, most nights, I don't have those problems.
If this sounds like something you would like to experiment with, then use The Brain to determine the exposure, but change the noise tolerance to say that you have a low tolerance for noise. The Brain should offer you a longer exposure.
Finally, circling back on my original topic entitled "Exposure's Effect on Color Quality," I still believe that longer exposures have an additional advantage. I don't see it in these two images, but I suspect the reason is that they are still too noisy. I am hoping that with more integration time, it will become clear.
Brian
EDIT: I chose not to run a de-noising filter on the images, nor did I enhance them in any other way except for a small black level adjustment. I wanted to show them in their raw, unadulterated form. They will look much better after running them through Affinity Photo.
At first, I had trouble processing these images of M51. I was following the same technique I use for star clusters, but I was not getting the result I expected. The image had a severe red bias. In a topic that I started a few days ago, entitled "Exposure's Effect on Color Quality," I speculated that the exposures I used were not long enough. I was wrong. The problem was with APP. It turns out that I need to run Star Color Calibration on galaxies when I use the Wratten #12, red, and green filters. Now that I have that behind me, I can move on with my analysis.
Looking at the attachment, your first reaction might be "I like the left-hand image best." I don't disagree with you. I like the HII regions in the arms (the red highlights.) The right-hand image has some red, but it is muted. The reason for that is due to the order of filters. East of the meridian, I start with red, then green, and finally Wratten #12. Red light wavelengths pass through the atmosphere more easily than green and blue at a given altitude above the horizon due to Atmospheric Extinction. The unfortunate consequence of starting with red is that the atmosphere is still unsteady after sunset. So, in the right-hand image, the atmosphere was turbulent when I captured reds. However, on the night before, the left-hand image, the atmosphere was steadier, and therefore red light was stronger.
If we overlook the loss of red highlights in the right-hand image, we can see that the galactic core and arms are brighter. How can this be? I spent 151 minutes on the left-hand image, but I got a worse result. The right-hand image took only 110 minutes, but it is so much brighter.
Oh no! There goes Brian again with his 300-second exposures. Let me say, that 300 seconds is completely arbitrary. I just wanted to make it significantly long enough so that the difference was obvious. Furthermore, I have a paltry 71mm aperture. A lot of you guys and gals have 8 inches and greater. You are scooping up far more photons than I. I don't doubt that you are saying to yourself, "Brian, that's a pretty cr**py image for 2 hours of effort." Yes, that's what I get for having a small aperture. Believe me, if I could start all over again, I'd go straight for an 8-incher.
The point that I want to make is that you can get a brighter image in less time if you increase exposure. And there is the proof. Now of course there could be many reasons why you don't want to go longer. Among those are: severe light pollution, poor guiding, dodging clouds, and aircraft. Those are all valid reasons. Thankfully, most nights, I don't have those problems.
If this sounds like something you would like to experiment with, then use The Brain to determine the exposure, but change the noise tolerance to say that you have a low tolerance for noise. The Brain should offer you a longer exposure.
Finally, circling back on my original topic entitled "Exposure's Effect on Color Quality," I still believe that longer exposures have an additional advantage. I don't see it in these two images, but I suspect the reason is that they are still too noisy. I am hoping that with more integration time, it will become clear.
Brian
EDIT: I chose not to run a de-noising filter on the images, nor did I enhance them in any other way except for a small black level adjustment. I wanted to show them in their raw, unadulterated form. They will look much better after running them through Affinity Photo.