As I'm sure you know, if the stars are oblong, there will always be less fine detail. If the system is blurring the stars into a other than round shape, then it is doing the same with the fine details as well.
I also did a similar comparison a while back, I shoot in Bortle 3-4 and the dark sky really works well with longer subs, but not too long. For instance, I have found out that 300s LRGB and 600s SHO gives me the most detail. Drop lower and I start losing fine detail, go higher and there is no gain in any detail, it actually seams there are worse results since external factors (such as wind) have higher impact on overall sub quality in longer exposures. I think the sky darkness as well as aperture size and speed are all three very important in deciding the sub frame length. Great video as always.
Nice job Luke, as always. My sense is that the shorter exposure / more frames stacked image has smoother noise, but the other seems to have slightly better colour AND slightly better contrast. I think the lesson for us all is to get a minimum of 20 frames to stack, with exposures that are as long as our gear AND environmental conditions will allow for that still deliver perfectly round stars. In other words, the answer to the question is, as always "it depends!!!"
I know you probably like the red border around your UA-cam thumbnails because it sets them apart from others but unfortunately it's the same colour as the red line that appears under thumbnails of videos once I've watched them and it would be easy for me to skip over a new video. I think it's the reason why not many people do it.
The way I see it. The more subs you take the lower the normalised noise floor of the image. As long as you expose long enough to bring fine details slightly above that noise floor, then your good. So then it becomes a balancing act. Too long exposures and you don't get as many subs and therefore have more noise but stronger signal. Too short exposures you get many subs so lower noise floor but a weaker signal so might lose details to the normalisation process. I personally find for emission nebulae 300s is enough, granted I'm using a very sensitive cooled camera. For bright wideband stuff like Orion, Galaxies etc, 60s or even 30s is enough. 60 would blow Orions core out for me. Awesome testing!!! Really interesting video
Thank you ever so much my friend!! I agree, if you're exposing enough to swamp read noise on your camera in each sub then you'll be golden, it's just a matter of balance as to how many subs you're willing to deal with after that I guess! 👍
One thing i think is a benefit with shorter exposures is that when seeing conditions are a bit poor as they often are in Florida’s west coast, the images tend to be a slight bit less blurred. The comparison between the faint galaxy in those pics looks similar to what can be experienced.
Glad you did the comparison. Besides avoiding data processing headaches, the only reason for longer subs is avoiding camera read noise. With a F2.2 scope from low Bortle 7 on a broadband target and the IMX571 sensor in high gain mode, I believe the read noise is swamped to less than 5% of total noise with anything over about 2 seconds. I have a spreadsheet on read noise calculations and so with 1 minute subs in your situation, you are still way way beyond camera read noise playing any material part in that test situation. I shoot 6-8 second subs on an F3.65 12” Unguided Dobsonian and so I’ve done the math to know when read noise raises it’s head and believe me, you are still not even close to having it matter. You could have done 10 min, 1 min, AND added 6 second (1/10th again) and other than the pain of stacking 1000 subs, you would not have had a read noise issue, but I bet the image would have been a bit more detailed still. (Oh, and I own PI for post processing but stack with SIRIL as it’s so blazingly fast)
Hey Steven!! I totally agree mate yeah, testing with such a fast scope, bright skies and low read noise camera leads to read noise itself being a none-issue in my subs, definitely I think more testing could be worthwhile with a slower scope to see if I can demonstrate some differences there :-) cheers!!
I do live stacking for public outreach at a local observatory. We live-stack and project to a large screen outside the obsevatory. We have an old, home-brew large (30”, F5) Newtonian, with rather poor tracking, so the exposures are limited to 20s max. We get great results with the short exposures. I can take the captured fits and post-process with PI for even better results. We do end up using higher gain that is ideal, but a few blown-out stars don’t bother me.
I recently tried pushing my humble 102mm scope as hard as I could by using a 1x flattener and 294m in bin1. The small well depth and small sampling of 0.67 meant I chose 30s exposures to combat seeing and guiding errors. I was ablt to throw out 25% of the exposures and ended up with a much better stack than I anticipated.
The one on the right definitely seems to have more detail and contrast ... at least to me. Maybe not when zoomed in on the tiny galaxy, but for sure when zoomed out all the way.
Hey mate, You are a right legend! Thank you for sharing your tips and experiences here at UA-cam-You sure are gifted at this universal hobbie we all share a strong affection for. Your talent and dedication is most appreciated and your presentations are extraordinary. May you continue such exceptional work well into the future buddy. Sending you many blessings and light from Australia!
@@lukomatico Awww . . . . Wow I wasn't sure my message would reach you but I am shocked it has. I just assumed you would receive too many comments to actually find time to read them! My adopted Mam is from Gateshead in UK, so it does help that I find similarity in you. I love how well spoken you are and how great you articulate your messages. There are a few exceptional communicators on UA-cam who provide insights on astronomy but you are the best of them all. Trevor Jones from Canada just visited Byron Bay here in Australia for an event called "Star Stuff". I like his presentations but his puppy is super annoying and it distracts from the quality of content. Blessings to you and yours, Annika xxxx
good comparison luke! i think the things like elongated stars and satellite trails aren’t a shortcoming of your comparison here, i feel like it’s an important thing to include because that’s exactly the risk you take with longer subs. less chance for dithering to help out, less of a loss if you lose a frame to a wind gust or guiding spike, et cetera. for those who want to dive deeper, Deep Sky Detail has a great video on Signal to Noise Ratio And Number of Subframes, i’d recommend checking it out for a better more fundamental understanding on how many and how long your subs should be! and clear skies luke!
Thanks ever so much Gabe! - I'm glad you see it that way my friend, I agree it adds a real world and valid variable to the experiment - as you quite rightly say, that's the risk you take with those long subs! :-) Clear skies my friend!
Really interesting. Thanks for taking the time to do this experiment. I've found that a sort of sweet spot is often around 3 minutes, especially with OSC. Star colors seems to be very decently perserved, and detail in the background doesn't seem to take a hit. Downside is that it will increase the number of individual subs by 8 pr hour. When you have projects often spanning 20+ hours, those extra subs all add up both in terms of processing time and storage. As an example using a camera that produces 51mb fits: - 20 hours of data using 300s subs = 240 files / 12gb - 20 hours of data using 180s subs = 400 files / 20,5 gb Of course... a way around that is to take say 15 minutes worth of 1m subs only for the stars.
The data totals can really get out of hand fast hey!! I've got a lot more planned for this, can't wait to get stuck in further and see what we find :-) Clear skies mate!!
What I found out is that in most cases it's enough to shoot 120-180 seconds shots at broadband (OSC) and 180-300 on Narrowband (like the L-Ultimate). Most of the times I shot 30 seconds shots for the stars. To many different exposure times, will make the processing traject longer :) Nice video by the way !!
Interesting comparison. I think the "the longer, the better" is just a legacy of the times where cameras had a LOT more noise and much lesser quantum efficiency. Personally i opted for the 2 min exposures, because I never was able see any benefit on going to 3 min, 5 min or 6 min exposures. Tried several times and always went back. That way if i really have to drop a picture its less exposure times trashed. More pictures also mean more dittering. Last but not least , even if your mount tracking is PERFECT, you should always consider the wobbling of the atmosphere ... if you have a lot of jet stream on the night, the start would wobble a lot, causing bigger/elongated stars. Here shorter exposures would help. I'm not an expert, and my rig is much cheaper than yours, but i think a more modern approach is not "the longer exposure possible until stars become elongated", but the opposite one: "as short exposures as possible until noise start to increase". I have a touptek 26000kpa ( the same you presented some time ago ), and 2 mins is PLENTY to keep the noise down.
agree..it comes down to throwing away a 1-2min sub versus a 5min or longer when they are utter trash. It proves the point that TOTAL integration time is the KEY
Same here. Losing a sub in a shorter sequence of long exposures vs a few shorter ones has always been a concern for me. I generally run around 3 minute exposures. For me that gives me a good balance of tracking vs quantity of subs for a night’s work. Takes a bit longer to process of course, but time I have.
This subject has been talked about many times…general consensus is limiting to around 300 seconds on narrowband targets and much less on broadband. Too many potential problems with longer exposures…ex wind…poor guiding ..clouds etc. and loss of subs
Hey Luke! Thanks for making this video! This really shows the pros and cons for short vs long exposures. I had been doing experiments and your observations align with what I have discovered. The to get the best of both worlds I have started taking very short exposures for detail and longer exposures to bring out the contrast and faint dust. The biggest advantage for shorter exposures is less reliance on your tracking, especially for smaller objects. Clear Skies mate!
I’m doing this too. My scope is a C8 Edge HD with a wedge to make it Equatorial. Unfortunately it’s tracking is a bit crude and anything longer than about 20 or 25 seconds leads to poor stars. I’ve tried guiding and sometimes it works okay but it’s tricky. My captures have been stacked in SharpCap and just saved as-is (basic EAA) and move on to the next target! It’s not top quality but I enjoy doing it this way as I’m rather impatient!
So many starlink satellites now unfortunately makes shorter exposures almost a necessity. I took 100 exposures of Markarian's Chain the other night and more than half had either an aeoplane or satellite trail 😒
Hey mate! :-) there's certainly a lot of satellites up there now hey! At least with our stacking algorithms now they don't really ever appear in finished images, maybe I'll have to do a video on that!
Hi Luke intresting experiment like you say not much in it both images look good , the only down side I can see is the amount of hard drive space and processing time the shorter sub would take up, would be interesting to see how this same experiment on the Quattro , thanks again Luke for taking the time to make these videos
Thanks ever so much for watching Tony mate, as always!! :-D I think doing more of these experiments would be useful for folks, so I'll probably do it! Cheers!
I completely understand your reluctance to upgrade your computer, and it feels like an unnecessary expense, On the other hand, for the price of a set of good second-hand 36 nm narrowband filters, you can significantly enhance your processing enjoyment. And take full advantage of the new IMX sensors and the more CPU/GPU intensive algorithms that get the best out of your hard-won photons... AP is not a cheap hobby. lol
I can agree with that yeah! For me generally speaking, upgrading my pc is a no-brainer as it gets far more use than my telescopes in terms of hours used, and I need it for my livelihood too, the processing speed and freedom is a nice boost on the side though hahaha, it'd be hard to go back now I've been spoiled! Cheers! :-D
Nice idea for an episode, thanks for sharing! For NB I like longer 5-10min and Broadband 2-3 min exposures. I find the cores of most galaxies. Start to get out of hand after three minutes.
Good comparison Luke, I would certainly be happy with either of those images. To be honest I'd never really given that much thought to this, I have just taken longer exposures if using narrowband filters and shorter ones when imaging broadband targets. I know that's not very scientific but looking at the various images that appeared in the media and on-line when I began taking an interest in astro photography that was the overall impression I was given about how other people were doing it. From your comparison I would say there really is hardly any difference, though I would tend towards the shorter length exposures simply because there appeared to be less noise.
Another stellar presentation Luke, thanks for doing this as its a question that has been lingering for a while now. The question has resurfaced now that I've started shooting in mono with SHO filters. From what I could see in the video, I think the added stacks give the shorter subs the edge in regards to minor detail. Not to mention the issue of guiding that many of us run into as you noted. I think there is a sweet spot that needs to be found with a minimum number of subs. I would definately love to see a similar test done with SHO as well as a rough outline of needed ration of Ha vs Sii vs Oiii. I've been going with a 1:2 ratio for Ha vs Oiii and 1:1.5 for Ha vs Siii, but I'm not sure if that's a good benchmark. Again - THANKS!
Thanks so much Rob!! More tests to come with this for sure 👍👍 On the note of dividing up your time for SHO, honestly I always used to just shoot 1/1/1! :-) worked for me!
Really interesting. I've recently switched to 1 min to 90 second subs as I need a LP filter to go further in my Bortle 4 skies and that makes everything come out with a blue hue that I find hard to remove. The other advantage of 60-90 second subs is that I don't need guiding, which keeps my set-up nice and simple (and yes, less satellite trails and lost subs when they're shorter). Thanks for doing this test. I am planning to buy blur and noise exterminator soon and will use your links.
That's really kind of you, thank you my friend!! I'd love to have a night or two imaging from bortle 4 skies, it'd be an interesting experiment! Cheers! 👍👍
Fascinating. I like how you showed both images at multiple stages. A variation to test can be doing the same comparison with a ccd on a slow scope. I think the difference will be drastic because of the much higher read noise associated with ccd sensors.
Thanks Luke for another great video. I love the comparison of the two exposure lengths. Both images are very nice. My way of thinking is that from a more light polluted sky setting the shorter exposures are going to be the winner. I'll have to test this out with my scope also. Thanks again and clear skies. Ray
Thanks so much Ray!! - I'd be interested to hear how your own experiments go my friend, I enjoy this kind of thing a lot!! :-) Hope you guys are doing well! 👍👍
I did a 10min test on Tarantula and compared it to my usual 5min subs, I can say I did collect a lot more data, but also collectted a lot more satellite trails as for in the Southern Hemisphere it is busy to the south, so the shorter exposures allow me to get clean subs without too many getting satellite trails. Was cool to see the target clearly without having to stretch the 10min sub.
Thanks so much for sharing your experience mate! I think I'll probably need to do an extra test on this kinda thing with a slower scope to see how much that skews things for me Cheers!
Excellent comparison Luke which shows the problem that most of us face in that longer subs may have a slight edge in some areas but, satellites and wind conditions can reduce the ability to take advantage of them. I have been pondering this myself as I'm now imaging from a bortle 3/4 area but on top of a hill in Scotland which is more often than not fairy windy. I'm currently designing my obsy and it's making me lean more towards a rotating dome style structure rather than a roll-off-roof to keep wind issues down to a minimum.
Thanks ever so much Pete mate!! - I really agree that a dome would be perfect if you're dealing with continual wind, I can only imagine how much that helps! Other than that, maybe a wind shield of some kind if you find your wind generally prevails from one direction, that might help put your scope in the wind's shadow too. Cheers mate!
I agree. The short exposures definitely show more (darker) stars. Again, this is to be expected with a higher S/N ratio. At one point, this should also affect details in DSOs.
Great job Luke. I always find myself debating exposure times. Seeing the differences are so…well, subtle, I think I can rest more easily knowing that whatever I do and whatever I get, it’s all going to be fine. It’s always good to go back to basics in this hobby.
Hi Luke. Assuming tracking/guiding wouldn't be a problem, and that you gather more than 15-20 individual frames, I would say that the longer subs, ironically (or maybe not) gave out a better star shape and star contrast. If you check out the stars present in the arms of the galaxy, I think you can see that they are sharper, as they contrast better with the spiral arms, and also, again maybe ironically, they are rounder. Perhaps blur x-terminator can do a better job in correcting the shape/PSF of the stars for a stack made up of longer exposures... just assuming... I think the real difference would be for narrow-band imaging, where fainter signal is simply not being picked up by the imaging sensor on shorter exposures, in comparison with longer ones, and also where background signal and camera noise play a much more important role in the final product. All in all, it is an excellent test for an ongoing debate: who was first ? the chicken or the egg ? :D All the best and good luck with the channel.
Thank you so much my friend!! I'm planning on doing a further test with narrowband for this reason 👍👍 (and perhaps some more with a slower scope too) Cheers!!
Interesting vid Luke. Thanks for taking the time to do this. Your viewers, and myself, might be interested to see the results from slower scopes. I also noticed you changed the gain and bias settings. This is a "black art" area most astro imagers fumble about with. It might be worthwhile making a vid how you set these. Again thanks
I think the longer yours subs, the more your mount and seeing will work against your final resolution. Thats why the 60 1 min stack resolves that tiny galaxy better. My motto? Dont go over 2 min with broadband but take lots of shorter subs.
Excellent video as always Luke. I would like to see this example demonstrated on a faint object with very low surface brightness, using NB and a long focal length, say f7 or longer e.g. OU4 (Squid Nebula). Let's see and example where the relatively few photons counted over the different exposure times really compete with the noise (The dark current noise generally increases at a slower rate than the light signal). In this video, the dark current noise is greatly drowned out by the local LP and half moon.
Hi Luke. Great vídeo. I am a traveller which usually means polar alignment every Session. I have also mostly favored toward shorter exposure as every minute you spend in polar alignment means less data and less signal. I Wonder If you could do a similar test but with the following two sets: 1) 60x1min, good polar alignment, qnd 2) 5x10min, excelent PA. The 10min diff would simulate the extra time one would spend on perfect PA.
Hey Ali! - interesting question mate! I always feel like getting your PA close is worth it, as long as you're not spending ages of course! It's perhaps something I could test eventually, thank you for the suggestion! 👍
Nice video as always, Luke ! Great comparisson, showcases nicely, that exposure time isn't really a factor anymore with these modern sensors. Read noise was always an argument for longer exposures, now you can tailor your exposure length to other parameters, like the amount of data your pc can handle, satelite trails elimination with stacking algorithms etc. I have settled at 180 sec exposures for OSC data and 300 sec for dual narrowband, works well in my bortle 5 skies and stacking doesn't kill my old pc. Clear skies Luke !
Fantastic video Luke. I love real experiments like this rather than BS opinions that people often spew out. You learn much more and there are always surprises. Cheers
Interesting comparison. I shot the Whirlpool myself last week, and I happened to go with about 60*3 minute subs, and there wasn't the world of difference between mine and either of yours. I might have a bit more detail in the dust lanes in 5195 in the final image, but then it's probably because I went a bit more extreme with BlurXTerminator. I think perhaps people sometimes obsess a bit much about the exposure time, and you need a comparison like this to show that it often doesn't matter at all in a lot of cases. But of course it does depend on what you are trying to capture too. It would be interesting to see a comparison where different sub times does make a big difference though, if there is such a case..
Thank you ever so much for the feedback and for sharing your personal experience too my friend!! - I agree by the way, it's likely something that bothers a lot of people and it probably shouldn't! I do want to repeat this test with a slower scope though, let's see what that shows! Cheers 👍👍
Great job on this one Luke! I personally stick between 1 and 5 minutes and usually average more like 2. That is mostly because I don't trust my mount all that much and it hurts me to think about throwing way 10 minutes of data due to trailed stars. I would love to see this test repeated with a scope around f/5 to see if there are any differences when you aren't using an ultra fast optic
Definitely due a testing session with slower optics mate yeah!! I'll get my next rasa test done with the dual band filter, and then remount a slower scope! 👍👍 Cheers!
Great one my friend!! That noise comparison we have to thank to Poseidon's performance, a monstrous camera with almost no noise, if you did that comparison with amp glow sensors that would be a huge win for the shorter side!! Clear skies man!!
Thanks Luke! I wonder if zwo did the same experiment when they were developing the seestar s50. I know their choice of 10sec exposures were mostly to get rid of the field rotation from it being alt/az but so far I'm pretty impressed with it for what it is. I managed to get a few subs of NGC 3621 in an attempt to get that new supernova, sn2024ggi from my precarious position on the peak of my roof. I was only willing to risk myself going up and down the ladder in the dark with the seestar with its lucomatico dew shield and a tablet. It did a lot better at getting images up there in the breeze than I thought it would. I'm at 39 degrees Latitude so it was pretty low on the horizon at about 16-18 degrees (hence, getting on the roof to get over the trees...). Now I just need to get off my backside and process those subs... All of that is to say that I worried about the exposure length being so short that I was wasting my time, so I'm glad to see you run this experiment with your experience... I'm just a beginning imager so thanks again, it helps! /Dave 👍👍
Hey Dave mate!! I've just gotta say wow, that's some dedication to the craft mate, getting up on your roof with ladders to take a shot, haha!! Respect mate 👍👍 Wishing you clear skies!!
Really interesting comparison Luke and the result was not what I was expecting. I suspect though it would be a win for the 10 minute subs over say a 3 hour session ? A future video on exposure lengths re a faint galaxy might be interesting ? I would guess that's where the 10 minute subs would shine but with the improvement in camera sensors that could well be wrong ? My only other suggestion for a future video would be a pixinsight comparison of an image processed entirely with default settings vs the other dialed in ? Anyway mate great video 👍
An advantage to more shorter subs that you don't mention is that if you're undersampled, you can't drizzle with only six subs, but sixty is more than enough. Also, the combination of drizzling and BlurXterminator, if you're even slightly undersampled, is better than BlurXterminator alone. Of course, if you had many hours of integration time (i.e., more subs), the longer subs could also take advantage of this too.
I would be interested in another test. I too am impressed with the latest software and, as a practitioner of EAA i like to keep my total exposures down to 30 mins so i can move on to other targets during an evening. I use 60s subs. I would be curious to see a 30min comparison vs your 60min vs something even longer to just see how well short runs can now match up with the “all nighters”… 👏👍
Great video! Interesting test, careful and fair comparison, fascinating result. Personally, I liked the increased contrast and color on the longer exposure, but you noted everything that I could see and reached a very fair conclusion. Your presentations are consistently impressive and the reason I subscribe!!
Overall, I would say that the longer exposures WOULD have been better, the limiting factor was definitely the tracking/stability errors because of the wind. I think you got a much better dynamic range in the longer exposures, however the detail was negated by the stability issue.
I try for ten minute subs and to get as many hours in as I can on a target but you do get to a point of diminishing returns. I shot some 7200 second subs on M57 in Ha, as it has a lovely Ha detail. The longest I have collected data on one target was Melotte 15, a total of 70 hours and 20 minutes.
I have found surprising results lucky imaging some DSO's with my 24" dob, getting a pleasing result of M42 in a bit over 2 minutes. This is an interesting topic for sure. Damo
Great comparison, Luke. I'll have to try some tests of my own with my ASI 2600mm pro/SW 100mm Esprit. I image from a Bortle 2 area but still seldom image over 3-minute sub for LRGB targets.
Thanks mate! - I found it absolutely worth the time to experiment on this, it'd be interesting to hear how your own experiment turns out if you give it a go! :-) Clear skies!
Great video, Luke. Would love to see one where the gain and offset is the same and the only differential is exposure time. P.s. If you're using OBS Studio (if that's what you're using) it's still a hair out of sync with the audio to video. Hit me up for a link if you don't know how to fix it.
Thanks mate!! - the recordings from OBS usually seem synced, same after export from davinci, but I think it slips a bit after upload to UA-cam sometimes! I could be wrong though, maybe I'm missing something 👍👍
Interesting comparison mate I tried longer subs with my RASA last year & got more data regarding dusty areas on galaxies but sacrificed colour by blowing it out so I generally don't take it any further than 3 minutes. Maybe I'll try again sometime,,,, clear skies
Hey Luke, it’s something I’ve been thinking about lately, but you’ve kinda made my mind up! I normally shoot 10 min subs, but have lost quite a few, due to aircraft and wind, so maybe dropping to 5 mins or even 3 would be better! One thing to think about though, have you got enough memory on the pc, my 2600 mm camera files are 50mb each regardless of sub length, so more subs means more space taken up on the pc, also I will have to do more calibration frames to match the new sub length (I’m lazy!) But still I think it’s worth doing! Thanks for sharing mate👍 Clear skies!
Thanks so much Simon mate!! - it's a balancing act isn't it haha, getting enough depth in your exposure, but not having 2500 frames to stack at the end of the night!! Can't wait to see what your own experiment turns up with this mate, I think some short subs will work out well! Cheers!! 👍👍
Another point in favor of shorter subs - if one of the 10 minute subs is bad for whatever reason (mount has a hiccup, big gust of wind, someone shines a light at the telescope) you've lost 10 minutes of the night. With the 1 minute subs, you've lost 1 minute of observing time. I like 3 minute subs, but that's only because with my cobbled together starter mount the best I could do was 30 seconds and 1/3 of the subs were flawed. I've got a nicer mount now, and by god I'm going to use it!
Excellent experiment, really well detailed explanation, I personally use between 2min and 5min exposure depending on the target, but you are daring to use 10min exposure with the RASA11 🥵, you are the master good job!!!!😅💪🏻
Great video! I've always wondered about this and at times struggle with the "correct" exposure times. So, I think what you are saying is that under perfect conditions - it's better to take longer subs?
Thanks mate! I think that if conditions were perfect and your keep-rate of subs was 100% then long subs are fine 👍 if you regularly lose a few percent of subs though then you're better off with shorter exposures, you'll make up the SNR difference and then some by being able to keep more of your total imaging time 👍
Great video Luke. Really interesting. I'm left wondering about mono now. So with a colour camera you found that the longer subs gave better colour but less detail. What would that mean for a LRGB setup? Where in effect colour is derived from the contrast and detail recorded in the Luminance channel or perhaps even the contrast and detail in all four mono channels. Looking forward to the next videos in this series.
Thanks so much mate, and that's an interesting question! - I imagine with mono this might be a very similar overall result, hopefully I'll get the chance to experiment with it! Cheers mate!
Very nice comparison, Luke! This seems to be an ever-present question among astrophotographers. It's honestly a minimal concern for me, and I hope I didn't miss this in your video, but was there a large difference in time for calibration and stacking in PI? I'm wondering about a session that have a lot more total exposure time and how each technique would 'stack up'. 🙂
Cutting off from the video a bit early… but some comments on what I did see. An auto-stretch is a very poor way to compare two datasets like this. Little shifts in light gradient, or aspects of the stacking artifacts you didn’t remove prior, will result in different auto-stretches. Deeper data may also be stretched more, resulting in a background that looks noisier when it may actually be cleaner. You’d need to take steps to normalize the two frames for visual comparison or, for a real analysis, use Subframe Selector. Also, when people thing about fewer longer subs vs more short subs, they’re usually thinking about under-exposure. But what you’ve done is engaged the dual gain stage on the camera vs native, where the dual gain stage is actually a bit more optimal in terms of signal vs noise. And you’ve well-exposed both examples. This would normally mean the shorter exposures produces the best signal vs noise in the final image, while the 0 gain option is a sane one because it still performs very well but allows you to use something like the RASA without an utterly insane amount of subs. For a slower telescope, you would almost always want to shoot it at the dual gain stage. If you wanted a good comparison for what people are usually considering, turn on the dual gain stage, capture a set of subs which are fully exposed (cutting shy of clipping desirable highlights, as you say) and compare that to a series of much shorter exposures. Like 2 minutes vs 30 seconds or whatever (maybe 0 gain is saner with the RASA if that’s all you have to compare). That will produce a very different set of parameters for the experiment. A couple gotchas on this experiment, too: the shorter subs suffer more with heavier light pollution as fainter signal is all the more mixed up with the light pollution, and also, a sensor which is extremely clean like you’re using is going to perform much better in this experiment at the short range than a messier one with higher sensor noise, banding, etc. You’d also want to make sure both datasets have enough subs for rejection algorithms to be efficient.
A lot of great points mate! - later in the video I do cover nonlinear comparison btw, but generally speaking I think with tests like this without investing extreme amounts of clear sky time I'll never be able to satisfy all viewers curiosities! Your experiment parameters would be equally valid, albeit a different experiment from mine - if I'd done it your way for example (which I considered) then someone would ask why I disadvantaged the short subs by not using an appropriately higher gain setting, just as an example :-D Hope that doesn't come across as antagonistic, just offering my thoughts from a creation perspective, Cheers! Luke
Doesn’t seem antagonistic at all. And I certainly appreciate the limited clear skies consideration. If you’re aware of the differences between what you did and the example I have, the feedback I have would be to cover that in the comparison video, as it is certainly not obvious to folks. And your experiment is also quite specific to these modern dual-gain sensors and will behave quite differently on many older sensors. Not to indicate what “should” be done on your channel, of course. Just some thoughts for consideration toward future content. The topic of individual exposure time relative to signal vs noise does seem to be something people ask about a lot.
Hi Luke, I assume that you did not drizzle for the 60x short exposures, right? It would make sense for a fair comparison. But on the other hand, if you did, it would be another big advantage for short exposures? Keep going with your great videos!
I wished I actually had the time to test 10s, 1min, 3 min and 5min exposures. For 10s we know thanks to Astro Biscuit that the davabtage is that you can do basically lucky imaging and get better resolution not having to chase the seeing. But you need a sky with very low light polution. Next new moon I'm hoping to go to a Bortle 3 sky and shoot some OSC BB and NB. For NB I'll shoot 5min exposures but for BB I don't know if I'll go with 10s to not clip the galaxies of Markarians Chain or go longer.
Hey mate! The average exposure lengths that you personally use for the bulk of your imaging will influence what you feel is 'long' or 'short' of course, they're loose descriptors by nature. I think 60s is almost certainly widely accepted as short exposures still - imagine for example you told other people at a star party that 'I'm taking long exposures tonight' they'd be expecting 300s+ subs, not 60s (and certainly not 11s, going from your example of the long exposure cutoff being
@lukomatico I understand, but there is a good reason to use very short exposures. People who do planetary photography do it all the time. Only use the sharpest sub-frames from all the ones you captured. This technique is also applicable for deep sky; I even go as short as 3 seconds for a sub-frame when doing deep sky. I sort the frames based on FWHM and reject all frames with FWHM above a certain value. With this technique, you are able to make far sharper images than with traditional "long" exposures. You only need a camera with a low read noise, but modern cameras do have a very low read noise with high gain
@@PeterK6502 Hey Peter! Ahh I'd love to do some more deep-sky lucky imaging again, I usually use sharpcap and let it live-stack sub-stacks for me while filtering for only lower FWHM subs at the same time 🙂 I think it could be an interesting comparison video! All the best, CS! Luke
Nice job, but I think your comparison is not very fair because we have to combine at least 25-30 subs to minimize many types of noise. It would be great if you will compare 3-4 hrs total exposure. 24*10min and 40*1min. Thx.
That's fair to say mate! 👍 I think without them they're equally disadvantaged, but maybe it would change the results in an unexpected way - thanks for the suggestion!
Very interesting experiment. My own conclusion is 'compromise'. Nice calm night go for five minute subs, not so calm go for one to three minute subs. If it's really breezy . . . watch Star Trek reruns.
Long time no speak Luke (sorry mate, have drifted away from astro to bird photography). I'd be curious to see 5 min subs from a dark sky site with low to no winds.
Hey there my friend! I usually determine offset values by measuring the histogram of looped bias frames in something like sharpcap, but for this I just used some previously known good settings 👍 Hope that helps!
Hi, I did the math on your set up and your "ideal" sub length in Bortle 7 at f2 and 300nm (visual spectrum) bandwidth is just 0.9 seconds. You could effectively do lucky imaging so you might like to try it again one evening with 2, 5 or 10 second subs vs. 1 minute over a 15 minute period. If you can do an hour great but you're going to have a LOT of stacking :-) The data below if for a 2600MC as I don't have your camera details to hand but let me know if you'd like me to run them and I plug 'em in. ---------------- Environment ---------------- Star Magnitude: 11.0 Background magnitude: 18.30 Bortle: 7.0 Seeing FWHM: 3.0 arcsec ----------------- Exposure ----------------- Sub Exposure: 60 secs Number of subs: 60 Total Integration Time: 60.0 mins ------------------ Filter ------------------ Filter Bandwidth: 300 nm Wavelength: 550 nm ---------------- Noise Calcs ---------------- Background signal: 45.4 e⁻/sec/px Electrons per sub: 2,723 e⁻/px Sub background noise: 52.2 e⁻/px Sub thermal noise: 0.04 e⁻/px Sub read noise: 1.25 e⁻/px Sub Total noise: 52.20 e⁻/px Stack Total noise: 404.35 e⁻/px One long sub total noise: 404.23 e⁻/px Extra noise: 0.03 % Ideal @ 5%: 0.3 secs Ideal @ 2%: 0.9 secs ----------- Telescope: EdgeHD 11 ----------- Type: Reflector Focal Length: 559 mm Aperture: 279 mm Focal Ratio: 2.0 Central Obstruction: 95 mm Total area: 542.2 cm² -------------- Telescope Calcs -------------- Image scale: 1.39 arcsec/px Field of View: 2.41 x 1.61 deg Airy Disk: 2.68 µm Airy Disk anglular FWHM: 0.42 arcsec Airy Disk linear FWHM: 1.13 µm Dawes limit: 0.50 arcsec Rayleigh limit: 0.50 arcsec Niquist limit: 2.78 arcsec ---------------- Pixel Data ---------------- Flux from star: 1.435e-15 W/m²/nm Star flux (telescope): 1.292e+06 photons/px Star signal: 1.176e+06 e⁻/px Total FWHM: 3.0 Central pixel (Seeing): 17.1 % Central pixel: 16.8 % Central pixel signal: 198,100 e⁻ Central pixel ADUs: 835,400 Central pixel Value: 1.00 Signal noise (cp): 445 e⁻ Background flux: 2,993 photons/px Background signal: 2,723 e⁻/px Background noise: 52.2 e⁻ Centre pixel noise: 454 SNR: 436 ---------- Camera: ASI 2600MC Duo ---------- Type: Colour Resolution: 6,248 x 4,176 Size: 23.5 x 15.7 mm QE: 91 % Pixel size: 3.76 µm Bin: 1 Image Scale: 1.39 arcsec/px Gain: 100 Temperature: -10 °C Read noise: 1.25 e⁻ Full well: 15,811 Dark current: 0.00070 e/s/px eGain: 0.23717 e/ADU
Thank you for that John!! I think it would be interesting to do some kind of live-stacking test for those short subs in order to reduce the insane amounts of data being generated to do it conventionally, -perhaps that could be an additional video all of its own! Thanks mate!
@@lukomatico great video! i was also thinking (fellow Rasa11 owner) to compare to Lucky Imaging approach, to try to get more resolution with trying to beat Seeing. FWIW Sharpcap has well rounded built in Live stacking, where it stacks good FWHM it in memory and can save out every minute or so
Hey mate! Great suggestion, I'd tried that before and want to attempt it again, it works really well! What I was doing is letting sharpcap stack up around 10m of exposure with on-the-fly rejection and then automatically save and restart a new stack, giving me effectively 10m 'subs' made of many shorter subs, then stacking them all together! Clear skies!! :-)
@@lukomatico Some other things you can do with f2 (I'm using a HyperStar so in the same boat), I often use gain=0 (or -25 on the ASI 2600 Duo). Even then an 8th magnitude star still saturates on OSC at 1 second (no filter). I wonder if you could whack up the gain to 300-400 and take 1/10th second exposures in a video and then stack them like you would a planet? You'll obviously lose a lot of the tonality of the nebula (which is superb at gain=0 but you'll get some impressive detail. Still thinking out aloud, perhaps the L channel with high gain luck-imaging-style exposures (combinations of the above) and R/G/B/H with gain=0 and 5-10 second exposures. If I ever saw a clear sky here in London I'd give it a go.
Very interesting video! And I appreciate you made this video too after the "integration time" one! I have the same sensor, IMX571C, but in Touptek sauce, what's the difference in the two gain values you mentioned? I only used 100 on 5mins subs on an F4 system, but probably should I change somthing? How can I determine the best value?
Thank you so much mate! Regarding gain & offset settings, I'd start at whatever your driver preset 'deep sky' or 'hcg' settings offer to begin with, I'd honestly say I do the vast majority of my imaging at around or a little above the HCG enabling threshold and leave it at that for most things! :-) My best advice would be if your histogram is comfortably within the first quarter/third of the graph then don't worry much bud! 👍👍 Clear skies!
@@lukomatico at the moment I enabled the HCG, but I'm not sure it's effective at gain 100... From the graphs on the camera manual it seems that that maybe 177 or like that would be better as it lowers the read noise keeping high DR and fullwell. What is the HCG enabling threshold?
Great video Luke. Ive just started to drop my sub lenght to 2 mins. Seems to me like i get less noisy images. Plus you can get a couple of subs where the seeing had gone great. Longer subs maybe gets spread. Clear skies my friend
I’ve been wondering about this recently Luke so thanks for the video. Would love to see a comparison with the new Askar 120 if you have time. I take 3min subs but wonder after watching this if I’d be better going with 1 or 2 min? Not sure if the slower f7 scope makes a difference with the choice of time? Cheers Simon
Great video Luke! It brings to mind something I can’t get my head around. If I have sixty 1 minute subs, when they are stacked is the resulting image still a 1 minute exposure in reality? Simplistically, haven’t all the subs been added up and divided by sixty?
As you say, the 60 subs have just been added together - adding them is actually strictly the same as just leaving the shutter open for 60 minutes (assuming perfect tracking, infinite dynamic range, and zero read noise), which would also just keep adding electrons (converted from photons) to the pixel wells. The division by 60 at the end is just a convenient scaling/normalizations. You can go back and forth between your averaged image and your added image at will (assuming sufficient numerical precision), which proves they are actually perfectly equivalent to one another!
As I understand it mate by stacking multiple exposures we're increasing the signal to noise ratio itself, - after an hour's worth of exposure, however you divide it up, you've got an hour's worth of signal, and an hour's worth of noise. The signal from the dso/space is pretty constant per unit of time/subexposure, there will be some degree of flux but over many subs it should become accurate/robust - the noise in each shot is quite random though (aside from fixed-pattern noise), so that averages out and by comparison to the actual signal we want, falls behind - thus increasing signal to noise ratio. - I hope I haven't misunderstood what you're asking! Cheers 👍
All good information, thank you both. I guess what can’t get clear is, say a specific pixel on your image should have a value of 100 in the perfect exposure. When you take your subs, this pixel could be on the range of 90 to 110. As you stack the images it will get closer to the correct value of 100. If you too fifty subs, the value could be say 98 to 102, but not 5,000 (50 times 100) because dividing down by the number of subs. So at the end of the day, haven’t you just got the nearest perfect picture for your exposure length, and the image isn’t of length times number of subs? Am I over thinking it????
Great job Luke! I’m a little thrown off, why playerone vs Zwo 571 cameras have different readout charts. I thought 100gain was start of high gain mode? Should it be 125? I have Zwo what’s the difference
Hey mate! The start point for HCG is slightly different on those two manufacturers that's all 👍👍 stick to 100 on the ZWO, but 125 wouldn't do harm really if you'd already taken data at that
It certainly would have been a different experiment at that point yeah! I felt like I needed to use an appropriately adjusted gain setting for the shorter subs as it felt like the most fair way to do things to me 👍 Cheers!
Not sure if anyone has done it well not that I could find. Could you test f2 vs normal filters in both LP and DB filters. So the reason behind the test request is how well do they work or don't work for certain speeds. So no filter, LP and DB at f2, f5~6 and f10. (depending on what OTA's you have) I think it would be very interesting to see the difference in the 9 images or lack of maybe???
Do you think having a RASA 11" with a 26mp camera was the reason why the results were very close? Do you think a C8 SCT with a ASI071MC Pro or maybe a ASI294MC Pro would show similar results between a 1 minute and 10 minute, maybe not the quality of your results, but in the context of those two cameras. Thanks
I think the speed of my scope played a large part in this for sure, I'm willing to continue testing with a slower scope once my tests are finished with this one though!! 👍👍
Thanks for the video. I do prefer doing one minute exposures to reduce the chance of collecting "Elon-ized" subs, although storage does become an issue. I'd like to see a similar comparison done for something like a 150 to 200mm aperture f/5 to f/8 system. Is it possible the star elongation was due to flexture between the OTA and your guide scope?
Thanks so much my friend! I'll certainly do more of these tests and share 👍👍 Re: flexure, it should be totally rigid on this rig as everything is metal-to-metal connections, usually I can get away with 10m subs fine but this night had occasional guests - great suggestion though, cheers!
As I'm sure you know, if the stars are oblong, there will always be less fine detail. If the system is blurring the stars into a other than round shape, then it is doing the same with the fine details as well.
Absolutely spot on mate, yep!! 👍
Looks more like it got blurred due to seeing
I also did a similar comparison a while back, I shoot in Bortle 3-4 and the dark sky really works well with longer subs, but not too long. For instance, I have found out that 300s LRGB and 600s SHO gives me the most detail. Drop lower and I start losing fine detail, go higher and there is no gain in any detail, it actually seams there are worse results since external factors (such as wind) have higher impact on overall sub quality in longer exposures. I think the sky darkness as well as aperture size and speed are all three very important in deciding the sub frame length. Great video as always.
So many great points mate, thank you for sharing your experience!! 👍👍
Clear skies!
Bonita la toma de la galaxia remolino buen procesado
Thank you my friend!
Thanks for giving over a very rare clear night to do this for your viewers. Much appreciated
My pleasure mate!! It's always fun to share an experiment like this 👍👍
Nice job Luke, as always. My sense is that the shorter exposure / more frames stacked image has smoother noise, but the other seems to have slightly better colour AND slightly better contrast.
I think the lesson for us all is to get a minimum of 20 frames to stack, with exposures that are as long as our gear AND environmental conditions will allow for that still deliver perfectly round stars. In other words, the answer to the question is, as always "it depends!!!"
Excellent summary!! :-) 👍👍
I know you probably like the red border around your UA-cam thumbnails because it sets them apart from others but unfortunately it's the same colour as the red line that appears under thumbnails of videos once I've watched them and it would be easy for me to skip over a new video.
I think it's the reason why not many people do it.
The way I see it. The more subs you take the lower the normalised noise floor of the image. As long as you expose long enough to bring fine details slightly above that noise floor, then your good. So then it becomes a balancing act. Too long exposures and you don't get as many subs and therefore have more noise but stronger signal. Too short exposures you get many subs so lower noise floor but a weaker signal so might lose details to the normalisation process. I personally find for emission nebulae 300s is enough, granted I'm using a very sensitive cooled camera. For bright wideband stuff like Orion, Galaxies etc, 60s or even 30s is enough. 60 would blow Orions core out for me. Awesome testing!!! Really interesting video
Thank you ever so much my friend!! I agree, if you're exposing enough to swamp read noise on your camera in each sub then you'll be golden, it's just a matter of balance as to how many subs you're willing to deal with after that I guess! 👍
One thing i think is a benefit with shorter exposures is that when seeing conditions are a bit poor as they often are in Florida’s west coast, the images tend to be a slight bit less blurred. The comparison between the faint galaxy in those pics looks similar to what can be experienced.
I hear you mate! - I've observed similar results before too :-) Clear skies!
Glad you did the comparison. Besides avoiding data processing headaches, the only reason for longer subs is avoiding camera read noise. With a F2.2 scope from low Bortle 7 on a broadband target and the IMX571 sensor in high gain mode, I believe the read noise is swamped to less than 5% of total noise with anything over about 2 seconds. I have a spreadsheet on read noise calculations and so with 1 minute subs in your situation, you are still way way beyond camera read noise playing any material part in that test situation. I shoot 6-8 second subs on an F3.65 12” Unguided Dobsonian and so I’ve done the math to know when read noise raises it’s head and believe me, you are still not even close to having it matter. You could have done 10 min, 1 min, AND added 6 second (1/10th again) and other than the pain of stacking 1000 subs, you would not have had a read noise issue, but I bet the image would have been a bit more detailed still. (Oh, and I own PI for post processing but stack with SIRIL as it’s so blazingly fast)
Hey Steven!! I totally agree mate yeah, testing with such a fast scope, bright skies and low read noise camera leads to read noise itself being a none-issue in my subs, definitely I think more testing could be worthwhile with a slower scope to see if I can demonstrate some differences there :-) cheers!!
I do live stacking for public outreach at a local observatory. We live-stack and project to a large screen outside the obsevatory. We have an old, home-brew large (30”, F5) Newtonian, with rather poor tracking, so the exposures are limited to 20s max. We get great results with the short exposures. I can take the captured fits and post-process with PI for even better results. We do end up using higher gain that is ideal, but a few blown-out stars don’t bother me.
Crikey that newt must be a sight to behold!! I'd love a big dob myself someday, EAA must be a lot of fun with that aperture! 👍👍
I recently tried pushing my humble 102mm scope as hard as I could by using a 1x flattener and 294m in bin1. The small well depth and small sampling of 0.67 meant I chose 30s exposures to combat seeing and guiding errors. I was ablt to throw out 25% of the exposures and ended up with a much better stack than I anticipated.
The one on the right definitely seems to have more detail and contrast ... at least to me. Maybe not when zoomed in on the tiny galaxy, but for sure when zoomed out all the way.
Hey mate,
You are a right legend! Thank you for sharing your tips and experiences here at UA-cam-You sure are gifted at this universal hobbie we all share a strong affection for. Your talent and dedication is most appreciated and your presentations are extraordinary. May you continue such exceptional work well into the future buddy.
Sending you many blessings and light from Australia!
You're too kind mate!! :-D Thanks so much for watching and clear skies to you!!
@@lukomatico Awww . . . . Wow I wasn't sure my message would reach you but I am shocked it has. I just assumed you would receive too many comments to actually find time to read them! My adopted Mam is from Gateshead in UK, so it does help that I find similarity in you. I love how well spoken you are and how great you articulate your messages. There are a few exceptional communicators on UA-cam who provide insights on astronomy but you are the best of them all. Trevor Jones from Canada just visited Byron Bay here in Australia for an event called "Star Stuff". I like his presentations but his puppy is super annoying and it distracts from the quality of content. Blessings to you and yours, Annika xxxx
Thank you so much for that!! :-) you're too kind!
All the very best to you and your family,
Luke
good comparison luke! i think the things like elongated stars and satellite trails aren’t a shortcoming of your comparison here, i feel like it’s an important thing to include because that’s exactly the risk you take with longer subs. less chance for dithering to help out, less of a loss if you lose a frame to a wind gust or guiding spike, et cetera.
for those who want to dive deeper, Deep Sky Detail has a great video on Signal to Noise Ratio And Number of Subframes, i’d recommend checking it out for a better more fundamental understanding on how many and how long your subs should be! and clear skies luke!
Thanks ever so much Gabe! - I'm glad you see it that way my friend, I agree it adds a real world and valid variable to the experiment - as you quite rightly say, that's the risk you take with those long subs! :-)
Clear skies my friend!
Really interesting. Thanks for taking the time to do this experiment.
I've found that a sort of sweet spot is often around 3 minutes, especially with OSC. Star colors seems to be very decently perserved, and detail in the background doesn't seem to take a hit.
Downside is that it will increase the number of individual subs by 8 pr hour. When you have projects often spanning 20+ hours, those extra subs all add up both in terms of processing time and storage.
As an example using a camera that produces 51mb fits:
- 20 hours of data using 300s subs = 240 files / 12gb
- 20 hours of data using 180s subs = 400 files / 20,5 gb
Of course... a way around that is to take say 15 minutes worth of 1m subs only for the stars.
The data totals can really get out of hand fast hey!! I've got a lot more planned for this, can't wait to get stuck in further and see what we find :-)
Clear skies mate!!
What I found out is that in most cases it's enough to shoot 120-180 seconds shots at broadband (OSC) and 180-300 on Narrowband (like the L-Ultimate). Most of the times I shot 30 seconds shots for the stars. To many different exposure times, will make the processing traject longer :) Nice video by the way !!
Thank you my friend!! 👍👍
Interesting comparison. I think the "the longer, the better" is just a legacy of the times where cameras had a LOT more noise and much lesser quantum efficiency. Personally i opted for the 2 min exposures, because I never was able see any benefit on going to 3 min, 5 min or 6 min exposures. Tried several times and always went back. That way if i really have to drop a picture its less exposure times trashed. More pictures also mean more dittering. Last but not least , even if your mount tracking is PERFECT, you should always consider the wobbling of the atmosphere ... if you have a lot of jet stream on the night, the start would wobble a lot, causing bigger/elongated stars. Here shorter exposures would help.
I'm not an expert, and my rig is much cheaper than yours, but i think a more modern approach is not "the longer exposure possible until stars become elongated", but the opposite one: "as short exposures as possible until noise start to increase". I have a touptek 26000kpa ( the same you presented some time ago ), and 2 mins is PLENTY to keep the noise down.
I think that's a very fair summary of things mate, cheers!! 👍👍
Certainly it seems that the days of needing extreme long exposures are gone
Great video Luke. I tend to stick to shorter lengths - max 2 or 3 mins - only because of satellite trails, and other unknowns 😂
I think that's a wise way to go about it mate!! Thank you for watching 👍
agree..it comes down to throwing away a 1-2min sub versus a 5min or longer when they are utter trash. It proves the point that TOTAL integration time is the KEY
Same here. Losing a sub in a shorter sequence of long exposures vs a few shorter ones has always been a concern for me. I generally run around 3 minute exposures. For me that gives me a good balance of tracking vs quantity of subs for a night’s work. Takes a bit longer to process of course, but time I have.
This subject has been talked about many times…general consensus is limiting to around 300 seconds on narrowband targets and much less on broadband. Too many potential problems with longer exposures…ex wind…poor guiding ..clouds etc. and loss of subs
This is great. I have to learn to use PI, these look incredible. I'm really new to the hobby, thanks for the video.
Hey Luke! Thanks for making this video! This really shows the pros and cons for short vs long exposures. I had been doing experiments and your observations align with what I have discovered. The to get the best of both worlds I have started taking very short exposures for detail and longer exposures to bring out the contrast and faint dust. The biggest advantage for shorter exposures is less reliance on your tracking, especially for smaller objects. Clear Skies mate!
Thank you ever so much Dave mate!! Your recent works have been stunning buddy 👍👍
Interesting. I normally go 4 to 15 seconds in Sharp Cap. I get bored waiting for the next update.
That's totally fair mate! You gotta do what makes you enjoy things the most 👍👍 clear skies!
I’m doing this too. My scope is a C8 Edge HD with a wedge to make it Equatorial. Unfortunately it’s tracking is a bit crude and anything longer than about 20 or 25 seconds leads to poor stars. I’ve tried guiding and sometimes it works okay but it’s tricky. My captures have been stacked in SharpCap and just saved as-is (basic EAA) and move on to the next target! It’s not top quality but I enjoy doing it this way as I’m rather impatient!
So many starlink satellites now unfortunately makes shorter exposures almost a necessity. I took 100 exposures of Markarian's Chain the other night and more than half had either an aeoplane or satellite trail 😒
Hey mate! :-) there's certainly a lot of satellites up there now hey! At least with our stacking algorithms now they don't really ever appear in finished images, maybe I'll have to do a video on that!
Hi Luke intresting experiment like you say not much in it both images look good , the only down side I can see is the amount of hard drive space and processing time the shorter sub would take up, would be interesting to see how this same experiment on the Quattro , thanks again Luke for taking the time to make these videos
Thanks ever so much for watching Tony mate, as always!! :-D I think doing more of these experiments would be useful for folks, so I'll probably do it!
Cheers!
Processing time and data storage, overall data collection efficiency etc. they can sway me away from too short of a sub.
I completely understand your reluctance to upgrade your computer, and it feels like an unnecessary expense,
On the other hand, for the price of a set of good second-hand 36 nm narrowband filters, you can significantly enhance your processing enjoyment. And take full advantage of the new IMX sensors and the more CPU/GPU intensive algorithms that get the best out of your hard-won photons...
AP is not a cheap hobby. lol
I can agree with that yeah! For me generally speaking, upgrading my pc is a no-brainer as it gets far more use than my telescopes in terms of hours used, and I need it for my livelihood too, the processing speed and freedom is a nice boost on the side though hahaha, it'd be hard to go back now I've been spoiled!
Cheers! :-D
Nice idea for an episode, thanks for sharing! For NB I like longer 5-10min and Broadband 2-3 min exposures. I find the cores of most galaxies. Start to get out of hand after three minutes.
Thanks mate!! - I think that sounds like a really nice approach, a good balance! :-) Hope you're getting clear skies bud!
Good comparison Luke, I would certainly be happy with either of those images. To be honest I'd never really given that much thought to this, I have just taken longer exposures if using narrowband filters and shorter ones when imaging broadband targets. I know that's not very scientific but looking at the various images that appeared in the media and on-line when I began taking an interest in astro photography that was the overall impression I was given about how other people were doing it. From your comparison I would say there really is hardly any difference, though I would tend towards the shorter length exposures simply because there appeared to be less noise.
Thanks for sharing mate!! More to come on this subject for sure 👍👍
Another stellar presentation Luke, thanks for doing this as its a question that has been lingering for a while now. The question has resurfaced now that I've started shooting in mono with SHO filters. From what I could see in the video, I think the added stacks give the shorter subs the edge in regards to minor detail. Not to mention the issue of guiding that many of us run into as you noted. I think there is a sweet spot that needs to be found with a minimum number of subs. I would definately love to see a similar test done with SHO as well as a rough outline of needed ration of Ha vs Sii vs Oiii. I've been going with a 1:2 ratio for Ha vs Oiii and 1:1.5 for Ha vs Siii, but I'm not sure if that's a good benchmark. Again - THANKS!
Thanks so much Rob!! More tests to come with this for sure 👍👍
On the note of dividing up your time for SHO, honestly I always used to just shoot 1/1/1! :-) worked for me!
Really interesting. I've recently switched to 1 min to 90 second subs as I need a LP filter to go further in my Bortle 4 skies and that makes everything come out with a blue hue that I find hard to remove. The other advantage of 60-90 second subs is that I don't need guiding, which keeps my set-up nice and simple (and yes, less satellite trails and lost subs when they're shorter). Thanks for doing this test. I am planning to buy blur and noise exterminator soon and will use your links.
That's really kind of you, thank you my friend!!
I'd love to have a night or two imaging from bortle 4 skies, it'd be an interesting experiment! Cheers! 👍👍
Excellent experiment! Would love to see a narrowband version of this experiment.
Working on that my friend! :-D
Fascinating. I like how you showed both images at multiple stages. A variation to test can be doing the same comparison with a ccd on a slow scope. I think the difference will be drastic because of the much higher read noise associated with ccd sensors.
Thanks mate! - I think doing the test again on a slower scope would be a good idea yeah! I'll have to invest some time into it :-) Cheers!
Thanks Luke for another great video. I love the comparison of the two exposure lengths. Both images are very nice. My way of thinking is that from a more light polluted sky setting the shorter
exposures are going to be the winner. I'll have to test this out with my scope also. Thanks again and clear skies.
Ray
Thanks so much Ray!! - I'd be interested to hear how your own experiments go my friend, I enjoy this kind of thing a lot!! :-)
Hope you guys are doing well! 👍👍
I did a 10min test on Tarantula and compared it to my usual 5min subs, I can say I did collect a lot more data, but also collectted a lot more satellite trails as for in the Southern Hemisphere it is busy to the south, so the shorter exposures allow me to get clean subs without too many getting satellite trails. Was cool to see the target clearly without having to stretch the 10min sub.
Thanks so much for sharing your experience mate! I think I'll probably need to do an extra test on this kinda thing with a slower scope to see how much that skews things for me
Cheers!
@@lukomatico I tested with an f5 200mm reflector. Guiding was .43 that night, came out amazing..... except for satellites
AWESOME video, Luke. You addressed so many things I had been wondering for a very long time. Absolutely fantastic content my friend!
Thank you so much my friend, as always I appreciate your support a great deal!! 👍👍
Excellent comparison Luke which shows the problem that most of us face in that longer subs may have a slight edge in some areas but, satellites and wind conditions can reduce the ability to take advantage of them. I have been pondering this myself as I'm now imaging from a bortle 3/4 area but on top of a hill in Scotland which is more often than not fairy windy. I'm currently designing my obsy and it's making me lean more towards a rotating dome style structure rather than a roll-off-roof to keep wind issues down to a minimum.
Thanks ever so much Pete mate!! - I really agree that a dome would be perfect if you're dealing with continual wind, I can only imagine how much that helps! Other than that, maybe a wind shield of some kind if you find your wind generally prevails from one direction, that might help put your scope in the wind's shadow too.
Cheers mate!
Thanks for doing this. I see more stars on the Left... I see more resolution on the Left... stars appear brighter on the Left.
Thanks so much Mike!! I'll have to repeat this with a slower scope, let's see what we find!
I agree. The short exposures definitely show more (darker) stars. Again, this is to be expected with a higher S/N ratio. At one point, this should also affect details in DSOs.
Great job Luke. I always find myself debating exposure times. Seeing the differences are so…well, subtle, I think I can rest more easily knowing that whatever I do and whatever I get, it’s all going to be fine. It’s always good to go back to basics in this hobby.
Hey Joe! - I hope you've been doing well my friend, it's good to hear from you! :-) I totally agree, sometimes going back to basics is key - cheers!
Hi Luke.
Assuming tracking/guiding wouldn't be a problem, and that you gather more than 15-20 individual frames, I would say that the longer subs, ironically (or maybe not) gave out a better star shape and star contrast. If you check out the stars present in the arms of the galaxy, I think you can see that they are sharper, as they contrast better with the spiral arms, and also, again maybe ironically, they are rounder. Perhaps blur x-terminator can do a better job in correcting the shape/PSF of the stars for a stack made up of longer exposures... just assuming...
I think the real difference would be for narrow-band imaging, where fainter signal is simply not being picked up by the imaging sensor on shorter exposures, in comparison with longer ones, and also where background signal and camera noise play a much more important role in the final product.
All in all, it is an excellent test for an ongoing debate: who was first ? the chicken or the egg ? :D
All the best and good luck with the channel.
Thank you so much my friend!! I'm planning on doing a further test with narrowband for this reason 👍👍 (and perhaps some more with a slower scope too)
Cheers!!
Interesting vid Luke. Thanks for taking the time to do this. Your viewers, and myself, might be interested to see the results from slower scopes. I also noticed you changed the gain and bias settings. This is a "black art" area most astro imagers fumble about with. It might be worthwhile making a vid how you set these.
Again thanks
Thank you for the great feedback my friend, I appreciate it!! I'll try and incorporate some further tests for sure 👍👍
I think the longer yours subs, the more your mount and seeing will work against your final resolution. Thats why the 60 1 min stack resolves that tiny galaxy better. My motto? Dont go over 2 min with broadband but take lots of shorter subs.
Excellent video as always Luke. I would like to see this example demonstrated on a faint object with very low surface brightness, using NB and a long focal length, say f7 or longer e.g. OU4 (Squid Nebula). Let's see and example where the relatively few photons counted over the different exposure times really compete with the noise (The dark current noise generally increases at a slower rate than the light signal). In this video, the dark current noise is greatly drowned out by the local LP and half moon.
Hi Luke. Great vídeo. I am a traveller which usually means polar alignment every Session. I have also mostly favored toward shorter exposure as every minute you spend in polar alignment means less data and less signal. I Wonder If you could do a similar test but with the following two sets: 1) 60x1min, good polar alignment, qnd 2) 5x10min, excelent PA.
The 10min diff would simulate the extra time one would spend on perfect PA.
Hey Ali! - interesting question mate! I always feel like getting your PA close is worth it, as long as you're not spending ages of course!
It's perhaps something I could test eventually, thank you for the suggestion! 👍
Nice video as always, Luke ! Great comparisson, showcases nicely, that exposure time isn't really a factor anymore with these modern sensors. Read noise was always an argument for longer exposures, now you can tailor your exposure length to other parameters, like the amount of data your pc can handle, satelite trails elimination with stacking algorithms etc. I have settled at 180 sec exposures for OSC data and 300 sec for dual narrowband, works well in my bortle 5 skies and stacking doesn't kill my old pc. Clear skies Luke !
Clear skies my friend!! Definitely these modern sensors have changed the rules a little it seems! 👍
Fantastic video Luke. I love real experiments like this rather than BS opinions that people often spew out. You learn much more and there are always surprises. Cheers
So glad you enjoyed it my friend! It was a rather surprising one I think, I'm really glad I tried it out! 👍👍
More tests to come on this subject!
Interesting comparison.
I shot the Whirlpool myself last week, and I happened to go with about 60*3 minute subs, and there wasn't the world of difference between mine and either of yours. I might have a bit more detail in the dust lanes in 5195 in the final image, but then it's probably because I went a bit more extreme with BlurXTerminator.
I think perhaps people sometimes obsess a bit much about the exposure time, and you need a comparison like this to show that it often doesn't matter at all in a lot of cases.
But of course it does depend on what you are trying to capture too.
It would be interesting to see a comparison where different sub times does make a big difference though, if there is such a case..
Thank you ever so much for the feedback and for sharing your personal experience too my friend!! - I agree by the way, it's likely something that bothers a lot of people and it probably shouldn't!
I do want to repeat this test with a slower scope though, let's see what that shows!
Cheers 👍👍
Great job on this one Luke! I personally stick between 1 and 5 minutes and usually average more like 2. That is mostly because I don't trust my mount all that much and it hurts me to think about throwing way 10 minutes of data due to trailed stars. I would love to see this test repeated with a scope around f/5 to see if there are any differences when you aren't using an ultra fast optic
Definitely due a testing session with slower optics mate yeah!! I'll get my next rasa test done with the dual band filter, and then remount a slower scope! 👍👍
Cheers!
Great one my friend!! That noise comparison we have to thank to Poseidon's performance, a monstrous camera with almost no noise, if you did that comparison with amp glow sensors that would be a huge win for the shorter side!! Clear skies man!!
Thank you my friend!! 👍👍 I think it would be cool to test with other cameras and demonstrate the differences!
Clear skies bud!
Thanks Luke! I wonder if zwo did the same experiment when they were developing the seestar s50. I know their choice of 10sec exposures were mostly to get rid of the field rotation from it being alt/az but so far I'm pretty impressed with it for what it is. I managed to get a few subs of NGC 3621 in an attempt to get that new supernova, sn2024ggi from my precarious position on the peak of my roof. I was only willing to risk myself going up and down the ladder in the dark with the seestar with its lucomatico dew shield and a tablet. It did a lot better at getting images up there in the breeze than I thought it would. I'm at 39 degrees Latitude so it was pretty low on the horizon at about 16-18 degrees (hence, getting on the roof to get over the trees...). Now I just need to get off my backside and process those subs...
All of that is to say that I worried about the exposure length being so short that I was wasting my time, so I'm glad to see you run this experiment with your experience... I'm just a beginning imager so thanks again, it helps! /Dave 👍👍
Hey Dave mate!! I've just gotta say wow, that's some dedication to the craft mate, getting up on your roof with ladders to take a shot, haha!! Respect mate 👍👍
Wishing you clear skies!!
Really interesting comparison Luke and the result was not what I was expecting. I suspect though it would be a win for the 10 minute subs over say a 3 hour session ? A future video on exposure lengths re a faint galaxy might be interesting ? I would guess that's where the 10 minute subs would shine but with the improvement in camera sensors that could well be wrong ? My only other suggestion for a future video would be a pixinsight comparison of an image processed entirely with default settings vs the other dialed in ? Anyway mate great video 👍
Thanks so much Paul mate!! As always you have cracking suggestions 👍👍
I'll keep those in mind, definitely more tests to come on this one
Clear skies!
An advantage to more shorter subs that you don't mention is that if you're undersampled, you can't drizzle with only six subs, but sixty is more than enough. Also, the combination of drizzling and BlurXterminator, if you're even slightly undersampled, is better than BlurXterminator alone. Of course, if you had many hours of integration time (i.e., more subs), the longer subs could also take advantage of this too.
All very true mate, great points!! 👍👍
Interesting, thank you for taking the time to do this comparison and sharing the results.
My pleasure mate!! Cheers! :-)
I would be interested in another test. I too am impressed with the latest software and, as a practitioner of EAA i like to keep my total exposures down to 30 mins so i can move on to other targets during an evening. I use 60s subs. I would be curious to see a 30min comparison vs your 60min vs something even longer to just see how well short runs can now match up with the “all nighters”… 👏👍
Hey mate!! I have just such a video - hope this is interesting to you! ua-cam.com/video/hUhCIrIWfTU/v-deo.html
Great video! Interesting test, careful and fair comparison, fascinating result. Personally, I liked the increased contrast and color on the longer exposure, but you noted everything that I could see and reached a very fair conclusion. Your presentations are consistently impressive and the reason I subscribe!!
Thank you ever so much for that mate, cheers!! I'm glad to have your support :-)
Overall, I would say that the longer exposures WOULD have been better, the limiting factor was definitely the tracking/stability errors because of the wind. I think you got a much better dynamic range in the longer exposures, however the detail was negated by the stability issue.
That's very fair feedback mate, cheers!!
Im just blown away with how much detail you got with 1 hour of data. The Rasa is amazing. From this I would go for 2 or 3 min or maybe 5.
Thanks mate! - the RASA really is like cheat codes for astro data to a certain point, haha!
I try for ten minute subs and to get as many hours in as I can on a target but you do get to a point of diminishing returns. I shot some 7200 second subs on M57 in Ha, as it has a lovely Ha detail. The longest I have collected data on one target was Melotte 15, a total of 70 hours and 20 minutes.
Thanks for sharing that Jules!! - I bet those long m57 subs really revealed the outer structure, wonderful idea 👍👍
Clear skies mate!
Excellent video. Surprised me with the results.
Thank you David!! I want to repeat this with a slower scope and see how that fairs, could be interesting!
As always Luke, a solid walktrough processing images. Thanks mate👍
Thanks my friend!!
I have found surprising results lucky imaging some DSO's with my 24" dob, getting a pleasing result of M42 in a bit over 2 minutes. This is an interesting topic for sure.
Damo
Excellent stuff Damo! Clear skies mate 👍👍
I like them both... but I favor the left just slightly better. Nice video comparison, Luke👍
Thanks mate!! - it's a difficult choice haha! :-)
Great comparison, Luke. I'll have to try some tests of my own with my ASI 2600mm pro/SW 100mm Esprit. I image from a Bortle 2 area but still seldom image over 3-minute sub for LRGB targets.
Thanks mate! - I found it absolutely worth the time to experiment on this, it'd be interesting to hear how your own experiment turns out if you give it a go! :-) Clear skies!
I like them both... but I favor the left just slightly better. Nice video comparison Luke👍
Thank you very much mate!!
You've answered this age old question that I've had! Thank you!
Thank you my friend!!
Great video, Luke. Would love to see one where the gain and offset is the same and the only differential is exposure time. P.s. If you're using OBS Studio (if that's what you're using) it's still a hair out of sync with the audio to video. Hit me up for a link if you don't know how to fix it.
Thanks mate!! - the recordings from OBS usually seem synced, same after export from davinci, but I think it slips a bit after upload to UA-cam sometimes! I could be wrong though, maybe I'm missing something 👍👍
Interesting comparison mate I tried longer subs with my RASA last year & got more data regarding dusty areas on galaxies but sacrificed colour by blowing it out so I generally don't take it any further than 3 minutes. Maybe I'll try again sometime,,,, clear skies
More tests incoming, Clear skies my friend!! 👍👍
If we don't take short enough exposures we'll never have reason to go buy more hard drives :)
Very true mate, haha!! :-D
Hey Luke, it’s something I’ve been thinking about lately, but you’ve kinda made my mind up! I normally shoot 10 min subs, but have lost quite a few, due to aircraft and wind, so maybe dropping to 5 mins or even 3 would be better! One thing to think about though, have you got enough memory on the pc, my 2600 mm camera files are 50mb each regardless of sub length, so more subs means more space taken up on the pc, also I will have to do more calibration frames to match the new sub length (I’m lazy!) But still I think it’s worth doing! Thanks for sharing mate👍 Clear skies!
Thanks so much Simon mate!! - it's a balancing act isn't it haha, getting enough depth in your exposure, but not having 2500 frames to stack at the end of the night!!
Can't wait to see what your own experiment turns up with this mate, I think some short subs will work out well!
Cheers!! 👍👍
Another point in favor of shorter subs - if one of the 10 minute subs is bad for whatever reason (mount has a hiccup, big gust of wind, someone shines a light at the telescope) you've lost 10 minutes of the night. With the 1 minute subs, you've lost 1 minute of observing time. I like 3 minute subs, but that's only because with my cobbled together starter mount the best I could do was 30 seconds and 1/3 of the subs were flawed. I've got a nicer mount now, and by god I'm going to use it!
Very true mate, aye!! Those long subs carry a lot more risk of loss!
Enjoy that new mount my friend 👍👍
Excellent experiment, really well detailed explanation, I personally use between 2min and 5min exposure depending on the target, but you are daring to use 10min exposure with the RASA11 🥵, you are the master good job!!!!😅💪🏻
Thank you mate, haha! :-D I think 2-5m subs are perfect generally, nice choice!
Clear skies!
Great video! I've always wondered about this and at times struggle with the "correct" exposure times. So, I think what you are saying is that under perfect conditions - it's better to take longer subs?
Thanks mate! I think that if conditions were perfect and your keep-rate of subs was 100% then long subs are fine 👍 if you regularly lose a few percent of subs though then you're better off with shorter exposures, you'll make up the SNR difference and then some by being able to keep more of your total imaging time 👍
Very interesting, Luke. Nice work!
Thank you mate!
Very nice comparision Luke. Could you also try 1sec sub(lucky image) comparision with longer exposure?
I could certainly try shorter subs yeah! that could be fun :-D Cheers mate!
Great video Luke. Really interesting. I'm left wondering about mono now. So with a colour camera you found that the longer subs gave better colour but less detail. What would that mean for a LRGB setup? Where in effect colour is derived from the contrast and detail recorded in the Luminance channel or perhaps even the contrast and detail in all four mono channels.
Looking forward to the next videos in this series.
Thanks so much mate, and that's an interesting question! - I imagine with mono this might be a very similar overall result, hopefully I'll get the chance to experiment with it!
Cheers mate!
Very nice comparison, Luke! This seems to be an ever-present question among astrophotographers. It's honestly a minimal concern for me, and I hope I didn't miss this in your video, but was there a large difference in time for calibration and stacking in PI? I'm wondering about a session that have a lot more total exposure time and how each technique would 'stack up'. 🙂
Hey Greg! - there was a reasonable difference in processing time yeah, probably somewhere around 3-4x longer, maybe more!
Clear skies my friend!
Cutting off from the video a bit early… but some comments on what I did see.
An auto-stretch is a very poor way to compare two datasets like this. Little shifts in light gradient, or aspects of the stacking artifacts you didn’t remove prior, will result in different auto-stretches. Deeper data may also be stretched more, resulting in a background that looks noisier when it may actually be cleaner. You’d need to take steps to normalize the two frames for visual comparison or, for a real analysis, use Subframe Selector.
Also, when people thing about fewer longer subs vs more short subs, they’re usually thinking about under-exposure. But what you’ve done is engaged the dual gain stage on the camera vs native, where the dual gain stage is actually a bit more optimal in terms of signal vs noise. And you’ve well-exposed both examples. This would normally mean the shorter exposures produces the best signal vs noise in the final image, while the 0 gain option is a sane one because it still performs very well but allows you to use something like the RASA without an utterly insane amount of subs. For a slower telescope, you would almost always want to shoot it at the dual gain stage.
If you wanted a good comparison for what people are usually considering, turn on the dual gain stage, capture a set of subs which are fully exposed (cutting shy of clipping desirable highlights, as you say) and compare that to a series of much shorter exposures. Like 2 minutes vs 30 seconds or whatever (maybe 0 gain is saner with the RASA if that’s all you have to compare). That will produce a very different set of parameters for the experiment. A couple gotchas on this experiment, too: the shorter subs suffer more with heavier light pollution as fainter signal is all the more mixed up with the light pollution, and also, a sensor which is extremely clean like you’re using is going to perform much better in this experiment at the short range than a messier one with higher sensor noise, banding, etc. You’d also want to make sure both datasets have enough subs for rejection algorithms to be efficient.
A lot of great points mate! - later in the video I do cover nonlinear comparison btw, but generally speaking I think with tests like this without investing extreme amounts of clear sky time I'll never be able to satisfy all viewers curiosities!
Your experiment parameters would be equally valid, albeit a different experiment from mine - if I'd done it your way for example (which I considered) then someone would ask why I disadvantaged the short subs by not using an appropriately higher gain setting, just as an example :-D
Hope that doesn't come across as antagonistic, just offering my thoughts from a creation perspective, Cheers!
Luke
Doesn’t seem antagonistic at all. And I certainly appreciate the limited clear skies consideration.
If you’re aware of the differences between what you did and the example I have, the feedback I have would be to cover that in the comparison video, as it is certainly not obvious to folks. And your experiment is also quite specific to these modern dual-gain sensors and will behave quite differently on many older sensors.
Not to indicate what “should” be done on your channel, of course. Just some thoughts for consideration toward future content. The topic of individual exposure time relative to signal vs noise does seem to be something people ask about a lot.
Hi Luke,
I assume that you did not drizzle for the 60x short exposures, right? It would make sense for a fair comparison.
But on the other hand, if you did, it would be another big advantage for short exposures?
Keep going with your great videos!
Thanks so much my friend! You're right, I did not drizzle - everything was stacked exactly the same :-) cheers!!
I wished I actually had the time to test 10s, 1min, 3 min and 5min exposures.
For 10s we know thanks to Astro Biscuit that the davabtage is that you can do basically lucky imaging and get better resolution not having to chase the seeing. But you need a sky with very low light polution.
Next new moon I'm hoping to go to a Bortle 3 sky and shoot some OSC BB and NB. For NB I'll shoot 5min exposures but for BB I don't know if I'll go with 10s to not clip the galaxies of Markarians Chain or go longer.
Good luck for clear skies in the b3 location mate!! That could be amazing 👍👍
I thought it was a comparison between long vs short, but this is long vs very long exposures.
For me, short exposure is
Hey mate!
The average exposure lengths that you personally use for the bulk of your imaging will influence what you feel is 'long' or 'short' of course, they're loose descriptors by nature.
I think 60s is almost certainly widely accepted as short exposures still - imagine for example you told other people at a star party that 'I'm taking long exposures tonight' they'd be expecting 300s+ subs, not 60s (and certainly not 11s, going from your example of the long exposure cutoff being
@lukomatico I understand, but there is a good reason to use very short exposures. People who do planetary photography do it all the time. Only use the sharpest sub-frames from all the ones you captured. This technique is also applicable for deep sky; I even go as short as 3 seconds for a sub-frame when doing deep sky. I sort the frames based on FWHM and reject all frames with FWHM above a certain value. With this technique, you are able to make far sharper images than with traditional "long" exposures.
You only need a camera with a low read noise, but modern cameras do have a very low read noise with high gain
@@PeterK6502 Hey Peter!
Ahh I'd love to do some more deep-sky lucky imaging again, I usually use sharpcap and let it live-stack sub-stacks for me while filtering for only lower FWHM subs at the same time 🙂
I think it could be an interesting comparison video!
All the best,
CS!
Luke
Nice job, but I think your comparison is not very fair because we have to combine at least 25-30 subs to minimize many types of noise. It would be great if you will compare 3-4 hrs total exposure. 24*10min and 40*1min. Thx.
Excelente!!
Would have loved to have seen the SNR numbers on each master ?
Thank you!! - Sorry I didn't run the script, I should have done!! I'll do that for the next comparison, cheers for the suggestion!
I feel like there should be proper dark and flat processing of the images to compare rightfully.
That's fair to say mate! 👍 I think without them they're equally disadvantaged, but maybe it would change the results in an unexpected way - thanks for the suggestion!
Great job luke cheers mate.
Glad you enjoyed it mate, thank you!! :-D
Very interesting experiment. My own conclusion is 'compromise'. Nice calm night go for five minute subs, not so calm go for one to three minute subs. If it's really breezy . . . watch Star Trek reruns.
Now that's a good take on things, haha!! :-D clear skies to you my friend!
Long time no speak Luke (sorry mate, have drifted away from astro to bird photography). I'd be curious to see 5 min subs from a dark sky site with low to no winds.
Hi Lukomatico. Nice comparison video! I wonder how do you setup the offset value, a I couldn't quite follow the way to do it. Thanks
Hey there my friend! I usually determine offset values by measuring the histogram of looped bias frames in something like sharpcap, but for this I just used some previously known good settings 👍
Hope that helps!
Hi, I did the math on your set up and your "ideal" sub length in Bortle 7 at f2 and 300nm (visual spectrum) bandwidth is just 0.9 seconds.
You could effectively do lucky imaging so you might like to try it again one evening with 2, 5 or 10 second subs vs. 1 minute over a 15 minute period. If you can do an hour great but you're going to have a LOT of stacking :-)
The data below if for a 2600MC as I don't have your camera details to hand but let me know if you'd like me to run them and I plug 'em in.
---------------- Environment ----------------
Star Magnitude: 11.0
Background magnitude: 18.30
Bortle: 7.0
Seeing FWHM: 3.0 arcsec
----------------- Exposure -----------------
Sub Exposure: 60 secs
Number of subs: 60
Total Integration Time: 60.0 mins
------------------ Filter ------------------
Filter Bandwidth: 300 nm
Wavelength: 550 nm
---------------- Noise Calcs ----------------
Background signal: 45.4 e⁻/sec/px
Electrons per sub: 2,723 e⁻/px
Sub background noise: 52.2 e⁻/px
Sub thermal noise: 0.04 e⁻/px
Sub read noise: 1.25 e⁻/px
Sub Total noise: 52.20 e⁻/px
Stack Total noise: 404.35 e⁻/px
One long sub total noise: 404.23 e⁻/px
Extra noise: 0.03 %
Ideal @ 5%: 0.3 secs
Ideal @ 2%: 0.9 secs
----------- Telescope: EdgeHD 11 -----------
Type: Reflector
Focal Length: 559 mm
Aperture: 279 mm
Focal Ratio: 2.0
Central Obstruction: 95 mm
Total area: 542.2 cm²
-------------- Telescope Calcs --------------
Image scale: 1.39 arcsec/px
Field of View: 2.41 x 1.61 deg
Airy Disk: 2.68 µm
Airy Disk anglular FWHM: 0.42 arcsec
Airy Disk linear FWHM: 1.13 µm
Dawes limit: 0.50 arcsec
Rayleigh limit: 0.50 arcsec
Niquist limit: 2.78 arcsec
---------------- Pixel Data ----------------
Flux from star: 1.435e-15 W/m²/nm
Star flux (telescope): 1.292e+06 photons/px
Star signal: 1.176e+06 e⁻/px
Total FWHM: 3.0
Central pixel (Seeing): 17.1 %
Central pixel: 16.8 %
Central pixel signal: 198,100 e⁻
Central pixel ADUs: 835,400
Central pixel Value: 1.00
Signal noise (cp): 445 e⁻
Background flux: 2,993 photons/px
Background signal: 2,723 e⁻/px
Background noise: 52.2 e⁻
Centre pixel noise: 454
SNR: 436
---------- Camera: ASI 2600MC Duo ----------
Type: Colour
Resolution: 6,248 x 4,176
Size: 23.5 x 15.7 mm
QE: 91 %
Pixel size: 3.76 µm
Bin: 1
Image Scale: 1.39 arcsec/px
Gain: 100
Temperature: -10 °C
Read noise: 1.25 e⁻
Full well: 15,811
Dark current: 0.00070 e/s/px
eGain: 0.23717 e/ADU
Thank you for that John!! I think it would be interesting to do some kind of live-stacking test for those short subs in order to reduce the insane amounts of data being generated to do it conventionally, -perhaps that could be an additional video all of its own!
Thanks mate!
@@lukomatico great video! i was also thinking (fellow Rasa11 owner) to compare to Lucky Imaging approach, to try to get more resolution with trying to beat Seeing. FWIW Sharpcap has well rounded built in Live stacking, where it stacks good FWHM it in memory and can save out every minute or so
Hey mate! Great suggestion, I'd tried that before and want to attempt it again, it works really well! What I was doing is letting sharpcap stack up around 10m of exposure with on-the-fly rejection and then automatically save and restart a new stack, giving me effectively 10m 'subs' made of many shorter subs, then stacking them all together!
Clear skies!! :-)
@@lukomatico Some other things you can do with f2 (I'm using a HyperStar so in the same boat), I often use gain=0 (or -25 on the ASI 2600 Duo). Even then an 8th magnitude star still saturates on OSC at 1 second (no filter). I wonder if you could whack up the gain to 300-400 and take 1/10th second exposures in a video and then stack them like you would a planet? You'll obviously lose a lot of the tonality of the nebula (which is superb at gain=0 but you'll get some impressive detail. Still thinking out aloud, perhaps the L channel with high gain luck-imaging-style exposures (combinations of the above) and R/G/B/H with gain=0 and 5-10 second exposures.
If I ever saw a clear sky here in London I'd give it a go.
Very interesting video! And I appreciate you made this video too after the "integration time" one! I have the same sensor, IMX571C, but in Touptek sauce, what's the difference in the two gain values you mentioned? I only used 100 on 5mins subs on an F4 system, but probably should I change somthing? How can I determine the best value?
Thank you so much mate!
Regarding gain & offset settings, I'd start at whatever your driver preset 'deep sky' or 'hcg' settings offer to begin with, I'd honestly say I do the vast majority of my imaging at around or a little above the HCG enabling threshold and leave it at that for most things! :-)
My best advice would be if your histogram is comfortably within the first quarter/third of the graph then don't worry much bud! 👍👍
Clear skies!
@@lukomatico at the moment I enabled the HCG, but I'm not sure it's effective at gain 100... From the graphs on the camera manual it seems that that maybe 177 or like that would be better as it lowers the read noise keeping high DR and fullwell. What is the HCG enabling threshold?
Great video Luke. Ive just started to drop my sub lenght to 2 mins. Seems to me like i get less noisy images. Plus you can get a couple of subs where the seeing had gone great. Longer subs maybe gets spread. Clear skies my friend
Great to hear it's working out so well for you Rob!! Clear skies my friend :-D
I’ve been wondering about this recently Luke so thanks for the video.
Would love to see a comparison with the new Askar 120 if you have time. I take 3min subs but wonder after watching this if I’d be better going with 1 or 2 min?
Not sure if the slower f7 scope makes a difference with the choice of time?
Cheers
Simon
Thanks so much for the suggestion Simon!! I'll absolutely revisit this test with a slower scope, I'm very interested in that too 👍👍
@@lukomatico Cheers mate, look forward to that 👍🏻🔭🌌
Great video Luke! It brings to mind something I can’t get my head around. If I have sixty 1 minute subs, when they are stacked is the resulting image still a 1 minute exposure in reality? Simplistically, haven’t all the subs been added up and divided by sixty?
As you say, the 60 subs have just been added together - adding them is actually strictly the same as just leaving the shutter open for 60 minutes (assuming perfect tracking, infinite dynamic range, and zero read noise), which would also just keep adding electrons (converted from photons) to the pixel wells. The division by 60 at the end is just a convenient scaling/normalizations. You can go back and forth between your averaged image and your added image at will (assuming sufficient numerical precision), which proves they are actually perfectly equivalent to one another!
As I understand it mate by stacking multiple exposures we're increasing the signal to noise ratio itself, - after an hour's worth of exposure, however you divide it up, you've got an hour's worth of signal, and an hour's worth of noise.
The signal from the dso/space is pretty constant per unit of time/subexposure, there will be some degree of flux but over many subs it should become accurate/robust - the noise in each shot is quite random though (aside from fixed-pattern noise), so that averages out and by comparison to the actual signal we want, falls behind - thus increasing signal to noise ratio. - I hope I haven't misunderstood what you're asking!
Cheers 👍
All good information, thank you both. I guess what can’t get clear is, say a specific pixel on your image should have a value of 100 in the perfect exposure. When you take your subs, this pixel could be on the range of 90 to 110. As you stack the images it will get closer to the correct value of 100. If you too fifty subs, the value could be say 98 to 102, but not 5,000 (50 times 100) because dividing down by the number of subs. So at the end of the day, haven’t you just got the nearest perfect picture for your exposure length, and the image isn’t of length times number of subs? Am I over thinking it????
Great job Luke! I’m a little thrown off, why playerone vs Zwo 571 cameras have different readout charts. I thought 100gain was start of high gain mode? Should it be 125? I have Zwo what’s the difference
Hey mate! The start point for HCG is slightly different on those two manufacturers that's all 👍👍 stick to 100 on the ZWO, but 125 wouldn't do harm really if you'd already taken data at that
I think if you got a lot more of short and long exposures you would get better results with long
Do you feel the switch to low read noise mode helped even the odds for both sides? I feel without it the experiment might have been quite different.
It certainly would have been a different experiment at that point yeah! I felt like I needed to use an appropriately adjusted gain setting for the shorter subs as it felt like the most fair way to do things to me 👍
Cheers!
Not sure if anyone has done it well not that I could find. Could you test f2 vs normal filters in both LP and DB filters. So the reason behind the test request is how well do they work or don't work for certain speeds. So no filter, LP and DB at f2, f5~6 and f10. (depending on what OTA's you have) I think it would be very interesting to see the difference in the 9 images or lack of maybe???
Thanks for the suggestion mate!! That could be an interesting test if I get enough clear skies to make it viable 👍👍 I'll keep it in mind!
Cheers!
Great video, thank you!
Thanks mate!!
Thanks!
Thank you ever so much my friend!! Cheers!! :-D
Good video. Why do you use 100 and 200 for your offset levels? I leave mine at 20 or 30. Wondering if I’m doing something wrong.
Mine's just set a bit overkill mate, don't worry about it too much! :-D you'll be fine
Quick question : did you adjust sensibility/gain on the 1m exposure to "compensate" the shorter exposure time?
Yes indeed mate :-)
Do you think having a RASA 11" with a 26mp camera was the reason why the results were very close? Do you think a C8 SCT with a ASI071MC Pro or maybe a ASI294MC Pro would show similar results between a 1 minute and 10 minute, maybe not the quality of your results, but in the context of those two cameras. Thanks
I think the speed of my scope played a large part in this for sure, I'm willing to continue testing with a slower scope once my tests are finished with this one though!! 👍👍
Thanks for the video. I do prefer doing one minute exposures to reduce the chance of collecting "Elon-ized" subs, although storage does become an issue. I'd like to see a similar comparison done for something like a 150 to 200mm aperture f/5 to f/8 system. Is it possible the star elongation was due to flexture between the OTA and your guide scope?
Thanks so much my friend! I'll certainly do more of these tests and share 👍👍
Re: flexure, it should be totally rigid on this rig as everything is metal-to-metal connections, usually I can get away with 10m subs fine but this night had occasional guests - great suggestion though, cheers!
Thanks
Thank you ever so much mate!! That's really good of you 👍👍
Clear skies!