I've learned so much about PixInsight from your videos but this one is probably the best one yet. When I took your ideas and applied them to my very low quality image of M31 (1 hour of unfiltered subs from a Bortles 7 site) it went from vague blob in the mist to an actually decent recognizable image of a Andromeda. Can't wait to see how it works once I have time to collect some real data!
Excellent video, have been using HSV star repair with mixed results. The mask approach with starnet as shown in your video is another great tool in the box. Cheers :)
Wow! Thank you! This helped me a ton! Can't get star net to work , I'll figure it out. I used standard masks on my M51 , still learning, and thank you!
This was great, Shawn. I've always struggled with getting good star colors without having the background turn to crap. IDK why I never thought of masks! I used the GAME script to create a mask for my galaxy--I was shooting NGC3184, and it worked perfectly-I'd forgotten all about it but it's also great.
Shawn, my utmost gratitude for your step by step instructions on how to color pop your images. You’ve actually encouraged me coz of your video here. I’m good at collecting images just no good at processing them coz I was intimidated by masks & had no idea how they work, but you my brother in Astrophotography have made masks more understandable & simple where others have made them seem complicated. Since your video I’ve processed 3 images & I’m more motivated now than ever before & have more confidence in my processing skills. Thanks. Any Patreon Page? Let me know…
This video was especially helpful to me. I found out that I had little clue how to color saturate. I wasn't even using the essential masks. Using this video, I got improved results from an image that had problems. My RGB stack, and subsequent steps produced pink stars. Very difficult for a rank amateur to correct. I had to try just 1 color on curves at a time to somewhat return stars to a more realistic color. Of course, I have no idea what imbalance caused this in the first place. Shouldn't STF balance my channels in the first place? I have a long way to go, but your videos are a great resource.
Before combining the individual channels, use the HistorgramTransformation tool to see which channel has the strongest signal (peak is furthest to the right) and then use LinearFit to match the other two channels to that one. Then combine using ChannelCombination. Follow that up with one of the three color correction tools (I've been using the AutoColor script). That should give you an image with a good starting point. I picked up most (if not all these tips) from watching Shawn's videos, so all credit goes to him.
@@feraxks yeah, I tried that using the statistics process and linear fit. For me it didn’t work. It produced a deep blue/green that I couldn’t use. So instead I took the pink image and inverted it and ran SCNR. That got rid of magenta and produced a better yellow that nevertheless would still not correctly saturate.
Great tutorial, thank you. I also watched your lum sharpening video which was just fabulous. My question is which would you do first. Sharpen the image then colour pop or colour pop before sharpening. Your guidance would be appreciated.
Starnet was working great for me for several months. After one of the last PixInsight updates, it stopped functioning. Re-installation not helping. The files are there but Starnet doesn't appear on any menu. If Shawn or anyone has a solution, please let me know. Thanks!
hi, i use PI to, I use a orion 120st f/5 refactor, so it causes some chromatic aberration, do you now if there is any tool that I can use do take off that aberration? And is there any tool that shrinks stars, some times I have to use it hehe? Your videos are helping me alot, thanks.
Hello Leticia! Thanks for watching and the comment/question. I not sure what you're up against exactly, I'd have to see an image, but you could try split channels command to create three individual mono frames and reduce stars in the Blue channel before recombining. If its magenta colour you could try the reduce magenta tip I made a video for here czcams.com/video/EgJfichawXQ/video.html . Let me know if either of these help. Clear skies!
2 starnet tips: - if you ran it once on your _L image for either a starmask or a starless version, you can get the other without running starnet again just with PixelMath: starless_image = extracted_L - star_mask or the other way around. - there seems to be a newer version of starnet around that's based on python and a newer version of tensorflow (standalone tool, don't think it's possible to make PI plugins in python). With the right DLLs from NVIDIA you can enable CUDA acceleration on that for a very significant speed boost if you have an NVIDIA GPU. My CPU is no slouch (16 core Ryzen 3950X) but my aging GTX 1070 beats it by a factor 3 to 4. When working on high-res images and small stride it saves a lot of time. github.com/nekitmm/starnet
really want to use this. will save me alot of time. Got it installed and now shows up in processes. woot. extracted a luminance layer from rgb and applied starnet++ as you did the in the vid. Now am getting a "error checkpoint file not found!". any ideas how to resolved appreciated.
Hi there - I don't know off hand what would be the cause of that or the solution. I do know it's finicky about directories and path names. Be sure you've installed it correctly and restart PI and try again. Failing this, I'd contact the developer and ask. Hope you work it out! Cheers.
Tchalla are you using Linux? If so, good, it's so much faster! But to fix your issue there is a command you have to run in the process console to point PI to the right directory, the command you need is in the Starnet readme file.
Starnet is overlooking a (bright) star in NGC6888. This is an issue in further processing. in fact, it is the star at position 20h12m13.15s / +38°26'35.4" (J2000). (that's what Stellarium is telling me) Can you confirm this? Is there any kind of group where this can be reported to?
Hi Shawn. Been trying to get Starnet working in pixinsigt. Could you share what version of pi are you using? I can’t get pi to recognise starnet as a module. Thanks
Please consider subscribing to my channel for more videos! Clear skies! 🙂
Thank you for providing step by step instruction at a pace that I can follow without pausing every three seconds!
Of all the PI tutorials I've watched, yours are the most descriptive and even paced. The audio brings them over the top...Subscribed!
Just found your channel…all I can say is THANK YOU!!!!!
I've learned so much about PixInsight from your videos but this one is probably the best one yet. When I took your ideas and applied them to my very low quality image of M31 (1 hour of unfiltered subs from a Bortles 7 site) it went from vague blob in the mist to an actually decent recognizable image of a Andromeda. Can't wait to see how it works once I have time to collect some real data!
You just taught me how to make masks in Pixinsight! Thanks!
Excellent video, have been using HSV star repair with mixed results. The mask approach with starnet as shown in your video is another great tool in the box. Cheers :)
Another great video. You have certainly helped me get to grips with PI. Keep em coming👍
Excellent tutorial Shawn!,on colour saturation well explained and easy to follow will try later after work,thank you for sharing.CS and stay safe.
Another really useful tutorial. Yes, Starnet++ is so useful to making great masks! Easy way to put it all together.
This was SO INCREDIBLY useful!!! I've been using Starmask and some masks I make from L. But both of these are way more effective
Great video, and it has given me some great ways to create masks. Thanks!
Well done on this video Shawn. Thanks for all the tips. Cheers
Another excellent video Shawn , Thank you
Super clear video, thanks!!
Happy and grateful to be aboard ~ 3weeks. Real information and not a shill product placement to be seen.
Wow! Thank you! This helped me a ton! Can't get star net to work , I'll figure it out. I used standard masks on my M51 , still learning, and thank you!
Very helpful--thank you!
This was a very helpful video on masks for a newbie like me to PI. Great video Shawn!! Thanks!
Thanks Jeff! Glad it helped. Cheers.
Awesome Shawn!!
This was great, Shawn. I've always struggled with getting good star colors without having the background turn to crap. IDK why I never thought of masks! I used the GAME script to create a mask for my galaxy--I was shooting NGC3184, and it worked perfectly-I'd forgotten all about it but it's also great.
great info again..thanks Shawn.
Thanks Paul!
Nice and informative as always. You can just hit the square in curves you don't need to drag the triangle.
Awesome video!.
Great video, perfect👍
Shawn, my utmost gratitude for your step by step instructions on how to color pop your images. You’ve actually encouraged me coz of your video here. I’m good at collecting images just no good at processing them coz I was intimidated by masks & had no idea how they work, but you my brother in Astrophotography have made masks more understandable & simple where others have made them seem complicated. Since your video I’ve processed 3 images & I’m more motivated now than ever before & have more confidence in my processing skills. Thanks. Any Patreon Page? Let me know…
Hi Bob! Thanks for the comment and kind words. Glad to know my efforts have been helpful. My patreon page is www.patreon.com/visibledark
Thank you so much.
Brilliant.
great video!!!!
Thanks!
Another fine, well paced, informative presentation 👏👏👏Quick Question Shawn. Can star 'points' be added to stars using PI please?
This video was especially helpful to me. I found out that I had little clue how to color saturate. I wasn't even using the essential masks. Using this video, I got improved results from an image that had problems. My RGB stack, and subsequent steps produced pink stars. Very difficult for a rank amateur to correct. I had to try just 1 color on curves at a time to somewhat return stars to a more realistic color. Of course, I have no idea what imbalance caused this in the first place. Shouldn't STF balance my channels in the first place? I have a long way to go, but your videos are a great resource.
Before combining the individual channels, use the HistorgramTransformation tool to see which channel has the strongest signal (peak is furthest to the right) and then use LinearFit to match the other two channels to that one. Then combine using ChannelCombination. Follow that up with one of the three color correction tools (I've been using the AutoColor script). That should give you an image with a good starting point.
I picked up most (if not all these tips) from watching Shawn's videos, so all credit goes to him.
@@feraxks yeah, I tried that using the statistics process and linear fit. For me it didn’t work. It produced a deep blue/green that I couldn’t use. So instead I took the pink image and inverted it and ran SCNR. That got rid of magenta and produced a better yellow that nevertheless would still not correctly saturate.
Great tutorial, thank you. I also watched your lum sharpening video which was just fabulous. My question is which would you do first. Sharpen the image then colour pop or colour pop before sharpening. Your guidance would be appreciated.
Both would be near last processing steps. Tweaks. The order wouldn't really matter to much.
Hi Shaun, I am literaly binge watching the videos. Please may I know, if the CenA was captured using OSC or a Mono cam with filters?
Starnet was working great for me for several months. After one of the last PixInsight updates, it stopped functioning. Re-installation not helping. The files are there but Starnet doesn't appear on any menu. If Shawn or anyone has a solution, please let me know. Thanks!
can you do the same with starxterminator ?
hi, i use PI to, I use a orion 120st f/5 refactor, so it causes some chromatic aberration, do you now if there is any tool that I can use do take off that aberration? And is there any tool that
shrinks stars, some times I have to use it hehe? Your videos are helping me alot, thanks.
Hello Leticia! Thanks for watching and the comment/question. I not sure what you're up against exactly, I'd have to see an image, but you could try split channels command to create three individual mono frames and reduce stars in the Blue channel before recombining. If its magenta colour you could try the reduce magenta tip I made a video for here czcams.com/video/EgJfichawXQ/video.html . Let me know if either of these help. Clear skies!
Starnet2 now.
2 starnet tips:
- if you ran it once on your _L image for either a starmask or a starless version, you can get the other without running starnet again just with PixelMath: starless_image = extracted_L - star_mask or the other way around.
- there seems to be a newer version of starnet around that's based on python and a newer version of tensorflow (standalone tool, don't think it's possible to make PI plugins in python). With the right DLLs from NVIDIA you can enable CUDA acceleration on that for a very significant speed boost if you have an NVIDIA GPU. My CPU is no slouch (16 core Ryzen 3950X) but my aging GTX 1070 beats it by a factor 3 to 4. When working on high-res images and small stride it saves a lot of time. github.com/nekitmm/starnet
really want to use this. will save me alot of time. Got it installed and now shows up in processes. woot. extracted a luminance layer from rgb and applied starnet++ as you did the in the vid. Now am getting a "error checkpoint file not found!". any ideas how to resolved appreciated.
Hi there - I don't know off hand what would be the cause of that or the solution. I do know it's finicky about directories and path names. Be sure you've installed it correctly and restart PI and try again. Failing this, I'd contact the developer and ask. Hope you work it out! Cheers.
Tchalla are you using Linux? If so, good, it's so much faster! But to fix your issue there is a command you have to run in the process console to point PI to the right directory, the command you need is in the Starnet readme file.
Problem solve by typing this into console cd "C:/program files/Pixinsight/bin"
Starnet is overlooking a (bright) star in NGC6888. This is an issue in further processing.
in fact, it is the star at position 20h12m13.15s / +38°26'35.4" (J2000). (that's what Stellarium is telling me)
Can you confirm this?
Is there any kind of group where this can be reported to?
Good question! I don't know where to report that to. I'll see if I can find out and get back to you on it.
Thanks a lot Shawn. This will be used on my next image. Look up astrobalcony on Instagram 😉
Hi Shawn. Been trying to get Starnet working in pixinsigt. Could you share what version of pi are you using? I can’t get pi to recognise starnet as a module. Thanks
Hi, how can I download starnet on MacBook?
Starnet is now included with the latest version of PixInsight. Under process/mask generation/starnet