A lot of you have asked me to show the result of stacking ALL the data... vs stacking less. I've put it up here : ruclips.net/user/DylanODonnellcommunity :)
Totally agree. James Lamb did an interesting video on this. While adding more frames does improve SNR this is at the expense of losing detail if you start to include poor FWHM subs. I guess the guest idea is collect lots of data so you can reject the worst and still have enough to get good SNR when stacking. Nice final image BTW. Oh, and I’m another guitar player. Must be a thing linking astronomy and guitar playing!.
Brilliant video Dylan, first off I'm glad to hear you are OK! - secondly, very interesting approach! I always cull crappy frames, usually at the end of a night I'll just take a quick look through the previews and get rid of frames I deem sub par, but this is quite a data-backed approach! Thanks for putting this together :) Clear skies! (but if its clear in the day, wear your sunscreen! 😅)
I just took your advice on some old data and...brilliant! Definitely saw a better result. Thanks! Love the channel. (And as an Aussie ex pat in the US, enjoy the Aussie flavor of the content!)
*Choir music* - Subframe selector tutorial Thanks so much for this - just what I needed. Also, glad to hear your nose is okay, and that you also have persistent leaks in your observatory - but for different reasons.
Dylan, Thanks, this is really good advice. I just reprocessed M106 data I took last week. My original image was made of 3.5hrs of 240s exposures. Visually, all frames looked sharp with round stars and I was reasonably happy with the result. This morning, after watching your video, I went back to the same data and removed anything that didn't score over 900 in DSS. This took me down to about 1hr 20mins worth of data. I reprocessed using my normal workflow in Photoshop (colour balance, stretch, noise reduction, saturation and sharpen the core a bit). Although the final image had a weaker signal, the fine detail in the photo was much sharper. I can see structure in the galaxy core that is just not there in the original and my original image looks slightly out of focus by comparison. This means I need to capture some more data on the subject to bring the overall SNR back to where I was but I can see now what I will gain by gathering more data whereas up until now, I never really saw any improvement after about 2 hours worth, probably because I was just layering up fuzziness. This always made me hesitant about wasting clear sky time on the same subject over and over when I could be chasing something new.
I really wish I understood what the hell I was watching when I watch your content, but I’m learning all the time by watching your stuff. Loving your work fella. I had a Mrs Button who was my sewing teacher in school.
I really like the approach you described. This makes complete sense. Screw Mr Brown he didn’t know anything about modern processors or the bloatware that chokes it.
Beautiful shot! I like the idea of the two camera trick you’ve been using instead of screwing around with the filters. The star count thing seems so obvious, but I bet a ton of people don’t think of that and you just made so many people’s lives easier there. Thanks for sharing as always. See, age spots from being almost 40 - lol. Glad you’re okay and I’m sure that was one of those “thankful for life” moments of mortality appreciation before we go back to space dust. Have a good one mate 🇦🇺 🍻 🇺🇸 great music 🎶 too - found your tunes on Amazon Music Unlimited. Tell your wife thanks 🙏 too if she helped you out
Computer "experts"! When Pentiums first came out I was told by the so called computer expert not to waste my money on one. "You will never use all the power it has, you could run a small country with it". Not sure if his name was Brown or not. Another great video mate and gives many of us something to aspire to with excellent explanation of what, why and how. Keep up the great work.
I've found that adaptive weighted average stacking (a form of multipass pixel rejection) does a good job of doing exactly this, without actually having to select frames. I still throw away the crap frames, but the final fine selection is done per-pixel by the pixel rejection algorithm, and I'm always surprised at how robust that is. Doing it per-pixel has the benefit of using all the available data for the background, improving SNR mildly there, where it is always so hard to get as smooth as one would want. For the curious folks, AWA stacking does an initial pass to estimate the average once, then another pass rejecting outliers (think k-s rejection) and compute a refined average+variance, and then a final pass where the refined average and standard deviation from the second step are used to compute a per-pixel weight for each sub, "soft-rejecting" outliers (instead of hard rejecting them). It's quite powerful (and will use your CPU even more, further irritating Mr Brown). It can be made even more robust by adding more passes, but I've found those 3 passes to be good enough for most purposes.
@@ForaxX I use my own, cvastrophoto (in github), but I took the algorithm from papers referenced by DSS so at least DSS has it. I think PI also has it, it may call it something else, I think Averaged Sigma Clipping. Edit: Nope... Averaged Sigma Clipping is slightly different. I think IKSS scaling estimator is a similar thing, but accomplishing what the AWA paper describes with it requires playing with settings nontrivially.
@@klaussfreire SWarp (by Emmanuel Bertin) does those sorts of coaddition algorithms too. We used this in different imaging pipelines back in the day for science grade data. For the convenience of the end-user, I recommend the THELI pipeline (by Micha Schirmer). It makes use of the Bertin scripts/programs and is fairly manageble to use.
I have been doing all my stacking in DSS from day 1, and have tried all the algorithms, and done comparisons of each using the same data each time. I agree, this method produces far superior results to all other stacking algorithms (in DSS at least). I use the default setting though, which I believe is 5 passes. I still do a quick analysis first before stacking and remove any subs that have significantly higher FWHM or sky background than the others. I don't put too much into the "stars" data, for 2 reasons: 1. I dither, so the actual number of stars varies by frame, and 2. The number of stars changes depending what detection threshold you chose. This could be more informative if you chose a low threshold, but since that makes stacking more CPU intensive without any actual benefit, I set the detection threshold as high as I can where the program can still stack properly (usually giving 25-50 stars on each sub).
Dylan, love the Centaurus A pic! Being from 40N, I'll pretty much never be able to image that one, despite really wanting to. Maybe if I move farther south after I retire, since I doubt I'll be bringing a 30kg mount with me on any vacations. Keep up the great work, love watching your videos man.
Excellent video. So many people advocate just using tons of subs saying more is better and not looking into the quality of the subs. Glad you clarified this for me. Also glad you get to keep your nose.
THanks for this Dylan. This runs contrary to most other folks take on data. Usually more is better but can cause issues with processing i.e light pollution etc (not something you have an issue with until all those celebs start throwing parties). Would be good to see images from all the data, 75%, 50%, 25% etc. Also some variation across different cameras I suspect high noise vs low noise, different targets (being a northerner I am not familiar with this target but I guess being a galaxy it is quite bright?).
I think he will eventually, he's the king when it comes to this... though wouldn't mind it taking a couple weeks longer now after seeing those perfect stars... jealous!!!
Amazing to see these in depth tours of your processing steps, and helpfull tips! Really adds a lot of value to a better understanding of the background of s top notch astro photo like this one.
Seriously getting rid of satellite tracks is the least of our concerns, considering all the many things that can go wrong. Sorting by star FWHM is a great plan, combined with the local contrast and the number of stars....blown away at how big a difference this sorting does compared to just taking all the data!
Typical British Nobby here, 5 nights of shooting is like an entire season for us haha. Great high level video that explains the approach. I use a very similar approach (only differences are that I use a slightly different formula for weighting and I apply a condition for FWHM and eccentricity in the top box above where you enter the weighting formula) and get great results, it's definitely worth doing! Also, your final image is amazing thanks for sharing.
Great job Dylan! I think this is one of my favorites pics I have ever seen from you, the target itself is really nice and the result amazing. Thanks for sharing and teaching us 👌👌
Great final image Dylan! Just a quick note, the process to output the subframes also writes the weights to the FITS headers with the keyword you specify. Then you can use in Image Integration to give better images more weight when combining. This is why it takes slightly longer than just copying the files over.
Another great video and hopefully it is just reminding us all to use only the best subs. Remember garbage in, garbage out. Glad the nose is froze, but man, that liquid N2 on your nose really can bring out the tears, eh? Sunscreen - you’re not rubbing it in, you’re rubbing it off. Is it great that the nights are getting longer? Not only the best stuff, but the best stuff with the most time. The Antipodes are so the place to be.
I second this comment. I’m actually working on a dataset that has about 25 hours of data in it and I used about 90 to 95% of it. I’m not super happy with the result, so I may try the Ruthless Method and do a side by side. Unfortunately, I don’t make videos about processing so unless you run across my astrobin account in about a week or so, you might never see the result.
Very nice article. You pointed out your EQ8-rh mount, are you still planning more videos covering the mount? I believe you mentioned something about polar alignment. I have ordered one but it is back ordered until December.
It’s always painful to toss those hard-won frames, but your approach makes a lot of sense. I find star count to be one of the most reliable ways to find the best frames. I use NINA and when the star count drops (during imaging) you know something has gone wrong.
Interesting, I wonder if PI handles it differently. I have been using DSS for stacking since day 1, and I usually ignore the star count because it's determined by the detection threshold in DSS, so I can change the number of stars just by changing that threshold and they tell you to set it low anyway because you only need a few dozen stars for proper alignment. More just increases CPU load and makes it take longer. Plus I dither, so the number of *actual* stars in each frame varies, even if all else is constant. Edit: just to be clear, I remove frames based on FWHM, and sky background (usually just outliers for this one, but more so if I know there were some clouds that night).
@@davidolszak6318 with NINA, the graph of the relative change in star-count (and in HFR) tell you something about changing conditions. Clouds will take down star-count, focus drift increases HFR, etc. It’s the relative change that I find useful. I would assume similarly useful (relative change) in PI SFS.
@@dmccallie Thanks for the reply! I don't get HFR readings in DSS, just FWHM and sky background (which is what I use to cull the lower quality subs). I could always register the subs with the threshold at 100, but that might cause it to also register hot pixels as stars. I'll look into NINA...I haven't researched it because I thought it was acquisition software and I use ASIair, not a PC.
Hi Dylan, great video as usual! Would you suggest the same logic to sorting R,G,B subframes when using a Mono camera and LRGB filters? Or only use this method for the Luminance channel?
Dr Robin Glover wants a word with you Dylan - more data is ALWAYS important in helping drive down thermal noise and random noise and improve the S/N ratio and thus preserve fainter data that is typically hidden by noise. edit: more QUALITY data. As Dylan does in this video, get rid of the sub standard data - OOF etc. Loved the video Dylan!
Hi! Great video!! One question: what is the cryteria you used in the choice by the number of the stars? i'm trying this on 3 night's lights on M81-M82. Thank you!!
I swear PixInsight is one of the most convoluted pieces of software I've ever seen. When I stack planetary I stack 1%, 2%, 3% etc up to about 40% and then go through by eye and work out the best one. Tedious but simple :P But hey if I got my pic on the Bintel FB page, the approach can't be that bad!
I agree...My goal is to extract as much data as I can with the least amount of time. My nights here are limited to weather and wildfires....so it is rare for me to have a string of good nights consecutively.
Great video. I am new to astronomy and am experimenting with altaz mount and smartphone with ISO set to 6400 and fast shutter, to reduce star trails. I took lots of pictures of the Beehive Cluster last night, which was so clear, and so will play with stacking today 😜😬👍🏼
Awesome image, m8! Glad your nose is healthy too. Yeah, living at an elevation of 8600 ' (2621 m), I've been zapped with liquid nitrogen many times despite using SPF50.
Really cool video. To select my frames I use the FWHM and SNR criteria in the subframeselector. I prefer using SNR instead of star count because if there is a cloud in front there will be less signal by default. The seeing conditions can also affect the signal quantity. All in all both methods are valid.
I once took 250 frames of RGB combined throughout an entire week of imaging and in the end only ended up using roughly 60 from across all nights. I could have used all of them but later dicided to only go for the best and image turned out amazing as a result. It was quite frustrating ignoring the vast majority of the data after 5 sleepiness nights but is was worth it in the end.
Cool vid again! Never used the subframe selector but will give it a try for sure. … BTW Is there an association as well between „Mr. Brown“ and Uranus? Just wondering.
Not even a lum/cut filter? That's pretty awesome. Really great photo- I think one of the best I've seen of this object. I only have access to it for a month every year here and it is LOW! Clear skies.
Gorgeous image Dylan. I just started doing what you do with the subframe selector myself but I was only getting rid of the very worst - maybe I have to be more bold in eliminating the lower quality image. Even getting rid of the worst had a profound effect. I noticed you don't use a luminosity filter, I have been using it because I have it. Do they actually do anything? I think there is some type of defect on mine that has been producing artifacts, fortunately I have been able to process it out. I was thinking on getting a new one but maybe just not use it. - Cheers
What timing! I was just last night looking at the SFS tool puzzling over how the damn thing works. I have almost 200 images of M5 that have the same quality issues your did...clouds and seeing-induced fuzziness, etc. This will let me pare that down to the best 50 or so. That is, if I can figure out the expressions and the best ones to try. To me they are, like everything else, meaningless. :-)
Dumb question maybe,but why are you not using the usb hub on the eq8 pro for the imagetrain ? i see allot of cables just hanging around,the usb hub will fix that all..right ? keep on going.
Nice description of the process. Definitely I will fit this thinking into my next project. But curious... did you find that the best images you ended up using were from one or two of the six nights mostly, or were they more distributed across the entire data set?
Definitely a mix.. the atmosphere is unpredictable enough that you'll get a soft shot on a good night or a suddenly good shot for a moment on a bad night.
Have you found you prefer the C11 over the RASA? I think I noticed you’ve been using that more lately. The idea of blocking part of the opening has always seemed a bit odd to me where you want a really small footprint up there, but I get the F/2 idea. Just curious since you’ve experienced both and get solid shots either way.
Thanks for the "nose" update and that it was good news . Anyway , great Video...I wish I was younger and smarter so I could wrap my head around PixInsight . That is some serious Software . I like your approach to "less is more" when it comes to the Data . People processing 27hrs of Data is just too much to comprehend . Cheers ! /SRK
I concur on sub count. I pick the best nights, no high clouds, not even a whisper. I also image 40 degrees up. I end up with 2 hours a piece of whichever filter I use, and I end up with exceptional detail. If I ever create a channel, ill share my work, but it is better than those who do 24 hours + of time.
Yep.. the worst nights are those where it *seems* clear but you just can't get those low HFR counts so you go over your gear a million times but ultimately, it's just the seeing.
Lucky location - we don't get many of those nights here in Tassie - there are always clouds moving in and out. Sometimes you are waiting for weeks for a clear sky night.
Lovely photo! Since you only stack the best frames, do you use a simple mean average for the highest SNR, or do you use something like sigma clipping? It would also be interesting to see the comparison of the auto stretch with all data included vs only the best frames.
I keep forgetting to tag you in my super crappy images, I hope you're not offended. Glad your nose isn't growing things it shouldn't! Gorgeous picture!
Poor Mr. Brown. He was probably wanting to leave a lasting impression; unfortunately, likely not about his analysis of computing power. Perhaps you can acquire an alpha version of a quantum or optical computing platform. Enjoyed the galaxy photo, but I have to say that the thumbnail star trails around the tree was quite nice, as was its larger full frame partner toward the end. Carry on, Dylan.
Great advice for managing data Dylan. Less is more, comes to mind ? Btw I'm very glad to hear your nose/skin problems weren't serious my friend! Wes, Liverpool, UK.
Hey Dylan I notice a copy of the DSM-5 on the bookshelf. Nowhere does it say anything about hawking Star Stuff to "Support my alcoholism..." And the question, "More than once gotten into situations while or after drinking that increased your chances of getting hurt (such as driving, swimming, using machinery, walking in a dangerous area, or having unsafe sex)?" doesn't mention astrophotography. Perhaps you're not an alcoholic but merely in denial about your lack of addiction. Of course, compulsive astrophotography might be a form of addiction. Let's see, have you spent a lot of money on astrophotography gear leaving you short of beer money? Do you spend a lot of time at night doing astrophotography instead of drinking? Have you spent time on astrophotography but not gotten the effects you wanted? Have you cut back on other activities, like drinking to do astrophotography? Have you wanted to get a certain astro-image so badly you couldn't think of anything else, even getting a cold one? Does doing astrophotography keep you away from family? Have been unable to cut back on astrophotography without suffering anxiety, perhaps necessitating a few brewskies? (Did you see what I did there?)
A lot of you have asked me to show the result of stacking ALL the data... vs stacking less. I've put it up here : ruclips.net/user/DylanODonnellcommunity :)
Me who has all bad data due to being untracked, unfiltered bortle 9 with a run of the mill mirrorless and a lens: **chuckles** this is fine
I must have been drunk and made another account and commented.... because this is me.
Phew, its not just me who tells their rig 'take this many' of the next object because I can't stay up any later, due to drunk... Hehe
Totally agree. James Lamb did an interesting video on this. While adding more frames does improve SNR this is at the expense of losing detail if you start to include poor FWHM subs. I guess the guest idea is collect lots of data so you can reject the worst and still have enough to get good SNR when stacking. Nice final image BTW. Oh, and I’m another guitar player. Must be a thing linking astronomy and guitar playing!.
Brilliant video Dylan, first off I'm glad to hear you are OK! - secondly, very interesting approach! I always cull crappy frames, usually at the end of a night I'll just take a quick look through the previews and get rid of frames I deem sub par, but this is quite a data-backed approach!
Thanks for putting this together :)
Clear skies! (but if its clear in the day, wear your sunscreen! 😅)
Thanks again Mr. O'Donnell, for another contribution to education.
thanks.. sometimes I say useful things :)
You just killed it, Dylan! Love the image. I use these in the subframe selector: In the approval button put this one FWHM
Great tip.. thanks Ray!
I just took your advice on some old data and...brilliant! Definitely saw a better result. Thanks! Love the channel. (And as an Aussie ex pat in the US, enjoy the Aussie flavor of the content!)
That’s great the final image is just awesome. I think I’m gonna go back and redo some of my images. Thanks man clear skies.
Your song is so catchy. Definitely adding it to my chill out playlists
Great video as always Dylan! Someday I’m going to visit the Southern Hemisphere so I can get that target and a few others.
*Choir music* - Subframe selector tutorial
Thanks so much for this - just what I needed. Also, glad to hear your nose is okay, and that you also have persistent leaks in your observatory - but for different reasons.
Nice shot Dylan 👍🏻
I realized not too long ago that the best data is better than a crap
load of data when I imagined the Whirlpool Galaxy.
Dylan, Thanks, this is really good advice. I just reprocessed M106 data I took last week. My original image was made of 3.5hrs of 240s exposures. Visually, all frames looked sharp with round stars and I was reasonably happy with the result. This morning, after watching your video, I went back to the same data and removed anything that didn't score over 900 in DSS. This took me down to about 1hr 20mins worth of data. I reprocessed using my normal workflow in Photoshop (colour balance, stretch, noise reduction, saturation and sharpen the core a bit). Although the final image had a weaker signal, the fine detail in the photo was much sharper. I can see structure in the galaxy core that is just not there in the original and my original image looks slightly out of focus by comparison. This means I need to capture some more data on the subject to bring the overall SNR back to where I was but I can see now what I will gain by gathering more data whereas up until now, I never really saw any improvement after about 2 hours worth, probably because I was just layering up fuzziness. This always made me hesitant about wasting clear sky time on the same subject over and over when I could be chasing something new.
Nice work!
I really wish I understood what the hell I was watching when I watch your content, but I’m learning all the time by watching your stuff. Loving your work fella. I had a Mrs Button who was my sewing teacher in school.
I really like the approach you described. This makes complete sense. Screw Mr Brown he didn’t know anything about modern processors or the bloatware that chokes it.
With your white coat you´re like Professor Proton for Astrophotography :D
Thank you Dylan for all the work and the inspiration. Clear nights!
Beautiful shot! I like the idea of the two camera trick you’ve been using instead of screwing around with the filters. The star count thing seems so obvious, but I bet a ton of people don’t think of that and you just made so many people’s lives easier there. Thanks for sharing as always. See, age spots from being almost 40 - lol. Glad you’re okay and I’m sure that was one of those “thankful for life” moments of mortality appreciation before we go back to space dust. Have a good one mate 🇦🇺 🍻 🇺🇸 great music 🎶 too - found your tunes on Amazon Music Unlimited. Tell your wife thanks 🙏 too if she helped you out
Hey thanks Joe!
Computer "experts"! When Pentiums first came out I was told by the so called computer expert not to waste my money on one. "You will never use all the power it has, you could run a small country with it". Not sure if his name was Brown or not. Another great video mate and gives many of us something to aspire to with excellent explanation of what, why and how. Keep up the great work.
I've found that adaptive weighted average stacking (a form of multipass pixel rejection) does a good job of doing exactly this, without actually having to select frames. I still throw away the crap frames, but the final fine selection is done per-pixel by the pixel rejection algorithm, and I'm always surprised at how robust that is. Doing it per-pixel has the benefit of using all the available data for the background, improving SNR mildly there, where it is always so hard to get as smooth as one would want.
For the curious folks, AWA stacking does an initial pass to estimate the average once, then another pass rejecting outliers (think k-s rejection) and compute a refined average+variance, and then a final pass where the refined average and standard deviation from the second step are used to compute a per-pixel weight for each sub, "soft-rejecting" outliers (instead of hard rejecting them). It's quite powerful (and will use your CPU even more, further irritating Mr Brown). It can be made even more robust by adding more passes, but I've found those 3 passes to be good enough for most purposes.
Super interesting! What software do you use? I don't remember this being implemented in PI
@@ForaxX I use my own, cvastrophoto (in github), but I took the algorithm from papers referenced by DSS so at least DSS has it. I think PI also has it, it may call it something else, I think Averaged Sigma Clipping.
Edit: Nope... Averaged Sigma Clipping is slightly different. I think IKSS scaling estimator is a similar thing, but accomplishing what the AWA paper describes with it requires playing with settings nontrivially.
@@klaussfreire SWarp (by Emmanuel Bertin) does those sorts of coaddition algorithms too. We used this in different imaging pipelines back in the day for science grade data. For the convenience of the end-user, I recommend the THELI pipeline (by Micha Schirmer). It makes use of the Bertin scripts/programs and is fairly manageble to use.
I have been doing all my stacking in DSS from day 1, and have tried all the algorithms, and done comparisons of each using the same data each time. I agree, this method produces far superior results to all other stacking algorithms (in DSS at least). I use the default setting though, which I believe is 5 passes. I still do a quick analysis first before stacking and remove any subs that have significantly higher FWHM or sky background than the others. I don't put too much into the "stars" data, for 2 reasons: 1. I dither, so the actual number of stars varies by frame, and 2. The number of stars changes depending what detection threshold you chose. This could be more informative if you chose a low threshold, but since that makes stacking more CPU intensive without any actual benefit, I set the detection threshold as high as I can where the program can still stack properly (usually giving 25-50 stars on each sub).
Dylan, love the Centaurus A pic! Being from 40N, I'll pretty much never be able to image that one, despite really wanting to. Maybe if I move farther south after I retire, since I doubt I'll be bringing a 30kg mount with me on any vacations. Keep up the great work, love watching your videos man.
Excellent video. So many people advocate just using tons of subs saying more is better and not looking into the quality of the subs. Glad you clarified this for me. Also glad you get to keep your nose.
Thanks Dylan, truly one of the most entertaining and informative channels on RUclips!!
Thx Ryan :)
THanks for this Dylan. This runs contrary to most other folks take on data. Usually more is better but can cause issues with processing i.e light pollution etc (not something you have an issue with until all those celebs start throwing parties). Would be good to see images from all the data, 75%, 50%, 25% etc. Also some variation across different cameras I suspect high noise vs low noise, different targets (being a northerner I am not familiar with this target but I guess being a galaxy it is quite bright?).
I can't comprehend why you don't have at least 100,000 subs.
Would be nice hah
I think he will eventually, he's the king when it comes to this... though wouldn't mind it taking a couple weeks longer now after seeing those perfect stars... jealous!!!
He has. Just Watch the Video
it's the intro... that loud, obnoxious intro! Has to be, otherwise the content is perfect!
Wait, what? Didn't we celebrate 1,000,000 subscribers recently? He lost that many in such a short time!?! :)
so underrated man, keep it up
👊🏼
Amazing to see these in depth tours of your processing steps, and helpfull tips! Really adds a lot of value to a better understanding of the background of s top notch astro photo like this one.
Glad it was helpful!
It's glorious, I'm astonished at the integration time and result. Very cool info too.
Seriously getting rid of satellite tracks is the least of our concerns, considering all the many things that can go wrong. Sorting by star FWHM is a great plan, combined with the local contrast and the number of stars....blown away at how big a difference this sorting does compared to just taking all the data!
Brilliant Dylan! That's so much smarter than what others do, stacking everything they've got.
Cracking image Dylan, so clear. Keep it up.
Your final picture is AMAZING!!!! great job Dylan!!
Glad to hear your ok dude. Great video as usual!!!
Typical British Nobby here, 5 nights of shooting is like an entire season for us haha.
Great high level video that explains the approach. I use a very similar approach (only differences are that I use a slightly different formula for weighting and I apply a condition for FWHM and eccentricity in the top box above where you enter the weighting formula) and get great results, it's definitely worth doing! Also, your final image is amazing thanks for sharing.
Haha nice work
Great job Dylan! I think this is one of my favorites pics I have ever seen from you, the target itself is really nice and the result amazing. Thanks for sharing and teaching us 👌👌
Thanks man :)
Sheff's kiss? Nah mate that image deserves a Mitchelin Star!
Great final image Dylan! Just a quick note, the process to output the subframes also writes the weights to the FITS headers with the keyword you specify. Then you can use in Image Integration to give better images more weight when combining. This is why it takes slightly longer than just copying the files over.
Another great video and hopefully it is just reminding us all to use only the best subs. Remember garbage in, garbage out.
Glad the nose is froze, but man, that liquid N2 on your nose really can bring out the tears, eh? Sunscreen - you’re not rubbing it in, you’re rubbing it off.
Is it great that the nights are getting longer? Not only the best stuff, but the best stuff with the most time. The Antipodes are so the place to be.
Incredible shot mate! The detail in the extended dust lanes is staggering. Great work! Also, love the BGM on this one :)
Thanks Neil :) that track will be up on Spotify etc soon !
Hey Dylan. Another awesome video. So happy to get the nose update!
Would be interesting to see the 4hr stack of everything to compare the two. Would really bang home the message!
I second this comment. I’m actually working on a dataset that has about 25 hours of data in it and I used about 90 to 95% of it. I’m not super happy with the result, so I may try the Ruthless Method and do a side by side. Unfortunately, I don’t make videos about processing so unless you run across my astrobin account in about a week or so, you might never see the result.
Here you go! ruclips.net/user/DylanODonnellcommunity
Excellent sorting technique. Very useful.
Very nice article. You pointed out your EQ8-rh mount, are you still planning more videos covering the mount? I believe you mentioned something about polar alignment. I have ordered one but it is back ordered until December.
"GOODSUBS" sounds like a great name for a new deli!
It’s always painful to toss those hard-won frames, but your approach makes a lot of sense. I find star count to be one of the most reliable ways to find the best frames. I use NINA and when the star count drops (during imaging) you know something has gone wrong.
Precisely !
I don't think I'll ever be able to throw so many frames away, no matter how much I try to convince myself it's better
Interesting, I wonder if PI handles it differently. I have been using DSS for stacking since day 1, and I usually ignore the star count because it's determined by the detection threshold in DSS, so I can change the number of stars just by changing that threshold and they tell you to set it low anyway because you only need a few dozen stars for proper alignment. More just increases CPU load and makes it take longer. Plus I dither, so the number of *actual* stars in each frame varies, even if all else is constant.
Edit: just to be clear, I remove frames based on FWHM, and sky background (usually just outliers for this one, but more so if I know there were some clouds that night).
@@davidolszak6318 with NINA, the graph of the relative change in star-count (and in HFR) tell you something about changing conditions. Clouds will take down star-count, focus drift increases HFR, etc. It’s the relative change that I find useful. I would assume similarly useful (relative change) in PI SFS.
@@dmccallie Thanks for the reply! I don't get HFR readings in DSS, just FWHM and sky background (which is what I use to cull the lower quality subs). I could always register the subs with the threshold at 100, but that might cause it to also register hot pixels as stars. I'll look into NINA...I haven't researched it because I thought it was acquisition software and I use ASIair, not a PC.
Love your presentation style.
Hi Dylan, great video as usual! Would you suggest the same logic to sorting R,G,B subframes when using a Mono camera and LRGB filters? Or only use this method for the Luminance channel?
Yesssssss Dylan is putting up pictures again. Beautiful shot.
Dr Robin Glover wants a word with you Dylan - more data is ALWAYS important in helping drive down thermal noise and random noise and improve the S/N ratio and thus preserve fainter data that is typically hidden by noise. edit: more QUALITY data. As Dylan does in this video, get rid of the sub standard data - OOF etc. Loved the video Dylan!
Hi! Great video!! One question: what is the cryteria you used in the choice by the number of the stars? i'm trying this on 3 night's lights on M81-M82. Thank you!!
Brilliant image, and very helpful process!
one of your best images, good work
I swear PixInsight is one of the most convoluted pieces of software I've ever seen.
When I stack planetary I stack 1%, 2%, 3% etc up to about 40% and then go through by eye and work out the best one. Tedious but simple :P
But hey if I got my pic on the Bintel FB page, the approach can't be that bad!
Hey Dylan, what's the app on your phone at 0:30 ? Great video! your ideas look like as if astrophotograpy had a revolution ;)
Sky guide for iOS
I agree...My goal is to extract as much data as I can with the least amount of time. My nights here are limited to weather and wildfires....so it is rare for me to have a string of good nights consecutively.
This is great! I love making CPU's do all the hard work.
'Music for Telescopes' ... very chill - like it! 👍🏽 ... and why not ... Brian May after all! Great tut on culling the sh**e. Thks!
Thanks :) That new track will drop into my artist page soon.
Great video. I am new to astronomy and am experimenting with altaz mount and smartphone with ISO set to 6400 and fast shutter, to reduce star trails. I took lots of pictures of the Beehive Cluster last night, which was so clear, and so will play with stacking today 😜😬👍🏼
Some tasty dust lanes there, high clouds be damned
Awesome image, m8! Glad your nose is healthy too. Yeah, living at an elevation of 8600 ' (2621 m), I've been zapped with liquid nitrogen many times despite using SPF50.
2:16 You scared me! I thought I had just bought something on the apple store by accident. LOL
when the beat kicks in at 9:00 boy it's so tasty, great soundtrack
Mastering track now .. should be on Spotify etc next week :)
@@DylanODonnell sick! Can't wait, excellent content, all-round, good stuff
Mr.Brown probably thought you would become a visual astronomer
Very good news about the nose! Excellent image of a galaxy I'll never get to see or shoot in person.
Really cool video. To select my frames I use the FWHM and SNR criteria in the subframeselector. I prefer using SNR instead of star count because if there is a cloud in front there will be less signal by default. The seeing conditions can also affect the signal quantity. All in all both methods are valid.
I once took 250 frames of RGB combined throughout an entire week of imaging and in the end only ended up using roughly 60 from across all nights. I could have used all of them but later dicided to only go for the best and image turned out amazing as a result. It was quite frustrating ignoring the vast majority of the data after 5 sleepiness nights but is was worth it in the end.
It pays to be picky! Speaking of, glad to hear your nose is ok!
Keep up the great work, so helpful... :)
Cool vid again! Never used the subframe selector but will give it a try for sure.
… BTW Is there an association as well between „Mr. Brown“ and Uranus? Just wondering.
Hehe
Not even a lum/cut filter? That's pretty awesome. Really great photo- I think one of the best I've seen of this object. I only have access to it for a month every year here and it is LOW! Clear skies.
Thx.. yep... no filter at all.. totally raw dog :)
@@DylanODonnell That's impressive- I've always had bad star bloat without one.
Gorgeous image Dylan. I just started doing what you do with the subframe selector myself but I was only getting rid of the very worst - maybe I have to be more bold in eliminating the lower quality image. Even getting rid of the worst had a profound effect. I noticed you don't use a luminosity filter, I have been using it because I have it. Do they actually do anything? I think there is some type of defect on mine that has been producing artifacts, fortunately I have been able to process it out. I was thinking on getting a new one but maybe just not use it. - Cheers
I guess I'm just a purist and like the idea of having the least amount of glass between me and the target. :)
@@DylanODonnell Well given your results I think I will turn into a purist as well - besides it's cheaper!
What timing! I was just last night looking at the SFS tool puzzling over how the damn thing works. I have almost 200 images of M5 that have the same quality issues your did...clouds and seeing-induced fuzziness, etc. This will let me pare that down to the best 50 or so. That is, if I can figure out the expressions and the best ones to try. To me they are, like everything else, meaningless. :-)
Brilliant Stuff man!! Love The Image!!
Well done! Awesome image!
100% on this. Since you posted this I have been doing the same thing, images are insane compared to stacking all the data
Yeh it’s really good huh? Shoot more use less!
Dumb question maybe,but why are you not using the usb hub on the eq8 pro for the imagetrain ? i see allot of cables just hanging around,the usb hub will fix that all..right ? keep on going.
cameras work best running straight to the computer and computer I have has LOTS of ports.. so no need to deal with usb hub issues :)
Yay! No cancer!!!! Great show!
Cliff Watson Pomona Astronomy Club, and see you on Dark Sky's show.
Thanks mate :) what dark skies show?
Lost in Light? I though that they said you were doing a talk. July 14? The Australian dark Sky Alliance.
Ahh thx :) didn’t realise they were advertising already !
Stunning image 👌. Good to hear about 👃. Poor Mr. Brown 😆
Nice description of the process. Definitely I will fit this thinking into my next project. But curious... did you find that the best images you ended up using were from one or two of the six nights mostly, or were they more distributed across the entire data set?
Definitely a mix.. the atmosphere is unpredictable enough that you'll get a soft shot on a good night or a suddenly good shot for a moment on a bad night.
Some amazing tips as usual, logical stuff that could easily be overlooked. Cheers mate! Did you use flats?
This time yes :)
Less is more...some good advice there Dylan! Thanks!
🤯 Amazing image Dylan!
Great video love it! I like the most how you express yourself about mister brown 🤣
Now that is an amazing image!
Awesome video dude. That image is so clean!
Have you found you prefer the C11 over the RASA? I think I noticed you’ve been using that more lately. The idea of blocking part of the opening has always seemed a bit odd to me where you want a really small footprint up there, but I get the F/2 idea. Just curious since you’ve experienced both and get solid shots either way.
I love them both! The c11 is my newest scope though so I've been enjoying shooting everything reeeeeally close :)
Thanks for the "nose" update and that it was good news . Anyway , great Video...I wish I was younger and smarter so I could wrap my head around PixInsight . That is some serious Software . I like your approach to "less is more" when it comes to the Data . People processing 27hrs of Data is just too much to comprehend . Cheers ! /SRK
thanks.. you too can join the cult of pixinsight if you want to :)
The master at work! Johannes Brahms apparently threw out more music than he published.
Hehe likely
I concur on sub count. I pick the best nights, no high clouds, not even a whisper. I also image 40 degrees up. I end up with 2 hours a piece of whichever filter I use, and I end up with exceptional detail. If I ever create a channel, ill share my work, but it is better than those who do 24 hours + of time.
Yep.. the worst nights are those where it *seems* clear but you just can't get those low HFR counts so you go over your gear a million times but ultimately, it's just the seeing.
Lucky location - we don't get many of those nights here in Tassie - there are always clouds moving in and out. Sometimes you are waiting for weeks for a clear sky night.
@@DylanODonnell Those are the nights to do widefield imaging at much less demanding pixel scales :) Bad seeing sucks, but can be worked around.
Lovely photo!
Since you only stack the best frames, do you use a simple mean average for the highest SNR, or do you use something like sigma clipping?
It would also be interesting to see the comparison of the auto stretch with all data included vs only the best frames.
Winsorized Sigma Clipping :)
Good afternoon in Portugal Dylan, for a focal length as long as the guiding camera you advise or use? Thanks
I’m using off axis guiding so the focal length is the same which works really well.
I keep forgetting to tag you in my super crappy images, I hope you're not offended. Glad your nose isn't growing things it shouldn't! Gorgeous picture!
Poor Mr. Brown. He was probably wanting to leave a lasting impression; unfortunately, likely not about his analysis of computing power. Perhaps you can acquire an alpha version of a quantum or optical computing platform. Enjoyed the galaxy photo, but I have to say that the thumbnail star trails around the tree was quite nice, as was its larger full frame partner toward the end. Carry on, Dylan.
Very sciency but easy to digest, love it!!!
I’m looking for a Musk filter to be sure, I order that my otherwise perfect frames don’t have the StarLink tracks...!
Thanks for the great pointers!
Awesome video mate, good job. I should really get PI one day, maybe next purchase ;)
Great advice for managing data Dylan. Less is more, comes to mind ? Btw I'm very glad to hear your nose/skin problems weren't serious my friend! Wes, Liverpool, UK.
Thx Wes :)
Thanks...was your first final image mono with no filters?
Correct :)
Glad to hear about the clean scan of the nose...oh, and the pic isn't bad either!
Hey Dylan I notice a copy of the DSM-5 on the bookshelf.
Nowhere does it say anything about hawking Star Stuff to "Support my alcoholism..."
And the question, "More than once gotten into situations while or after drinking that increased your chances of getting hurt (such as driving, swimming, using machinery, walking in a dangerous area, or having unsafe sex)?" doesn't mention astrophotography.
Perhaps you're not an alcoholic but merely in denial about your lack of addiction. Of course, compulsive astrophotography might be a form of addiction.
Let's see, have you spent a lot of money on astrophotography gear leaving you short of beer money?
Do you spend a lot of time at night doing astrophotography instead of drinking?
Have you spent time on astrophotography but not gotten the effects you wanted?
Have you cut back on other activities, like drinking to do astrophotography?
Have you wanted to get a certain astro-image so badly you couldn't think of anything else, even getting a cold one?
Does doing astrophotography keep you away from family?
Have been unable to cut back on astrophotography without suffering anxiety, perhaps necessitating a few brewskies? (Did you see what I did there?)
I feel like every new video your intro gets faster and more bass boosted
The early shows had a slower version of the same song which I improved at one point :)
DSM-5 on the bookshelf? Now that would take this channel in an interesting direction...
what sharp little eyes you have Mr bond.
After getting DSM-5 input, the next catchphrase will be: “Remember-we’re all crazy, everything is meaningless, and we’re all going to die.”
Wow, really sharp! It would have been very interesting to have also a version with 80% of the frames or so, for comparison.
Here you go :) Here you go! ruclips.net/user/DylanODonnellcommunity