5.73
#...
·
·
2
likes
|
|---|
|
I saw the original post shortly after it appeared and think there was but one reply at that time. My initial reaction was to reply right then, but then thought I should wait it out and see where it would go ... and it went just about where I thought it would go .... I agree with the OP about the "tools" many (most? / all?) of us use intending to better our results ~ there are obvious visible characteristics that signal overuse, but for people like me and @Kerry Bloor and @Dan Kearl (and perhaps hundreds - if not thousands - of others), with so little time to image, these tools are life savers. In my neck of the woods, it's not uncommon to have two imaging nights per dark cycle. Three is a huge luxury. It's looking like during this current dark cycle ~ 7-10 days before the New Moon and 5-7 days after ~ I will have one night to image (if the forecast is correct), and that means probably only 5-6 hours this month on one target ... but I can assure anyone who is reading this that I am very serious about creating top-quality images ... or at least, as best I can. Equipment aside, I just don't have the opportunity to acquire deeper integration on any target, though I long for it. Maybe someday .... To those of us who do have fantastic equipment under 250+ clear nights annually, AND have mastered the science and art of deep integration and awesome processing, keep your incredibly wonderful images coming. To the rest of us ~ irrespective of where we are on the path, let's keep on pressing forward. |
8.59
#...
·
|
|---|
Jay Hovnanian: FWIW, I have never had 250 clear nights. I don't think I've ever had 100 clear nights. Even if I had a quarter of the year worth of clear nights, I also could never take advantage of 100% of them. I do, however, try to take advantage of partially clear nights...which maybe is the key for other imagers who don't have a lot of 100% totally clear sky time. |
|
8.59
#...
·
|
|---|
Dan Kearl:Kerry Bloor: I think you've misundrstood. I'm definitely not trying to create anything like hallowed ground. I didn't make a comment that 30 hours of data was the minimum necessary either. Those are points I never made. A point I have tried to make is that I think with the advent of AI, there is too much reliance on it, and that AI processing is often taken to extremes. That instead of just taking AI to extremes, try to get more data. I guess that is overtly offensive now. A few posts back I asked what people think what a viable clear sky might be. I'm wondering if people are skipping nights that might be partially viable, at least some of the night...and if they instead took advantage of them, could that give them more opportunity to acquire data. That seems to have been offensive as well. This thread is now taking a much different turn, though, and I'm happy to delete it if people think its just going to be unnecessarily disruptive. I'm not trying to chase anyone off from posting, so don't stop posting on account of me. |
|
8.59
#...
·
·
1
like
|
|---|
Kerry Bloor: I am not quite sure what I'm incorrect about... I was more asking what people like you consider a viable "clear" night. I don't actually have a lot of time on my hands. That's where automation came in, as without it, I would have probably been out of the hobby long ago. I never disassemble my scope setup. I configured it so I could keep the scope, as well as a usb hub, rig runner and small PC as well as the vast majority of the cabling on the scope assembly. I then only had to worry about putting that whole assembly on the mount, and connecting two cables (one power, one data) to my power box. I would then head inside, program an imaging sequence for the night, make sure it took properly, then go back to doing all the other things I needed to. FWIW, I have no observatory, nothing at a remote imaging site, I don't rent time on dark sky telescopes, or anything like that. I have good, but not particularly exceptional equipment. Most of what allowed me to take advantage of as much clear sky as I could, even if its just holes in the clouds, was configuring the telescope with all the necessary stuff so I had a single "thing" to set up in just a couple of steps...and then automating the imaging sequence throughout the night. A LOT of my imaging was done with less than optimal equipment, the lens I used to use cost me a lot of data due to it being an image stabilized lens, and I spent most of my imaging time using a lower end mount as well. But I was still able to get some pretty good integration times. I think a lot of getting more data is not spending $100,000 on gear or hosting it at some remote observatory...but rather just minimizing the effort and steps to get going each night. Which is something that most people should be able to do, for most types of equipment. Anyway. |
1.51
#...
·
·
1
like
|
|---|
|
I can't claim to have read every word of the above conversation, but certainly most of it. So give me a pass if I missed something. I can remember Jon's posts from years ago. I, too, dropped out of the hobby for a few years, and when I returned last year, the impact of AI was underway. I have found many of these tools helpful since my focus is image processing. The whole process is so much easier now. I no longer collect data (we live in the cloudy NW), so I depend on data from others or buy it from sources like Telescope.live. I am often faced with creating a "decent" image from only a few hours of data. Plus, these data sets often have problems with the calibration process, or there are optical issues that sometimes result in severely distorted stars. Needless to say, I am very thankful for AI-developed tools like BlurXterminator and GraXpert AI. I agree there is no substitute for hours of quality data, but AI tools can often convert subpar data into an image that LOOKS okay, which is often good enough for me. This hobby attracts many types of people with varying goals. Given the limitations of the data I have available, the AI tools are a godsend. Tonight, I watched a recent interview with Jensen Huang, the CEO of Nvidea (https://www.youtube.com/watch?v=cEg8cOx7UZk&t=2467s). As you may know, Nvidea has revolutionized the development of machine learning. As I listened to him describe the impact of AI on many fields, I realized that image processing as we know it may not be long for this world. Imagine "Pixinsight AI," which could take the subs from an automated telescope/camera and produce an image many of us would consider acceptable. I'm not talking about a "crap" image like many computer-aided astronomy programs produce these days, but something that would be hard to distinguish from a quality image produced by today's methods. When I think about my processing steps, I do not think it would be that hard to train a computer to do what I do. This sort of development may be inevitable. Unfortunately, it may have a negative impact on processers such as myself, who still get a kick out of image processing. It could be the next step in the evolution of astrophotography. |
