Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Beginner Pixinsight tutorials?

This topic has been archived. This means that you cannot reply to this topic.
114 replies to this topic

#1 mostlyemptyspace

mostlyemptyspace

    Messenger

  • *****
  • topic starter
  • Posts: 444
  • Joined: 05 Jan 2014

Posted 12 January 2015 - 02:23 PM

Hi all. I normally do my AP processing in Photoshop after stacking in DSS. I keep hearing Pixinsight is the way to go, but it's pretty daunting. I looked up Pixinsight tutorials on the website, but they seem to be more advanced. I'm looking for some tutorials for a beginner that can walk me through an entire process. What do you guys recommend?



#2 Madratter

Madratter

    Hubble

  • *****
  • Posts: 13,277
  • Joined: 14 Jan 2013

Posted 12 January 2015 - 02:29 PM

Harry has some excellent video tutorials. They are what got me started.

You can try my PixInsight tutorial that takes you through processing some Messier 33 data. That is on my website.

#3 anismo

anismo

    Fly Me to the Moon

  • *****
  • Posts: 6,089
  • Joined: 30 Jan 2014

Posted 12 January 2015 - 02:36 PM

I also used lightvortexastronomy tutorials. http://lightvortexas.../tutorials.html

Very well done. You should also search in youtube. I found a bunch that I used when I was learning (and still am).



#4 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,974
  • Joined: 18 Jul 2011

Posted 12 January 2015 - 02:42 PM

I think the IP4AP ones are good (but not free) for learning how to use the different processes. Less so for a workflow.



#5 Goofi

Goofi

    Cosmos

  • *****
  • In Memoriam
  • Posts: 8,137
  • Joined: 03 May 2013

Posted 12 January 2015 - 04:54 PM

In addition to what the others are recommending (all good advice), what helped me the most was suggestions of scripts & processes to focus on.

 

BatchPreprocessing (a script) will register and integrate your subs

ScreenTransferFunction & HistogramTransfer (processes) will help you see your work

DynamicBackgroundExtraction is powerful, but I found ABE easier to use at first.

Noise Reduction ... pick one (or two) and learn it; many like TGVDenoise or MultiscaleMedianTransform

CurvesTransformation (a process) will help you tweak your image

 

It will help if you learn how to make a preview (Alt-N) and mask.

 

Best advice I got for learning was just take it one process at a time, go slow and as you get the hang of that process, add another.

It helps to think about what you're trying to accomplish, and then use the tool that lets you do that easily.

I'm sure others can weigh in with a list of the processes and scripts they like.



#6 mostlyemptyspace

mostlyemptyspace

    Messenger

  • *****
  • topic starter
  • Posts: 444
  • Joined: 05 Jan 2014

Posted 12 January 2015 - 07:17 PM

In addition to what the others are recommending (all good advice), what helped me the most was suggestions of scripts & processes to focus on.

 

BatchPreprocessing (a script) will register and integrate your subs

ScreenTransferFunction & HistogramTransfer (processes) will help you see your work

DynamicBackgroundExtraction is powerful, but I found ABE easier to use at first.

Noise Reduction ... pick one (or two) and learn it; many like TGVDenoise or MultiscaleMedianTransform

CurvesTransformation (a process) will help you tweak your image

 

It will help if you learn how to make a preview (Alt-N) and mask.

 

Best advice I got for learning was just take it one process at a time, go slow and as you get the hang of that process, add another.

It helps to think about what you're trying to accomplish, and then use the tool that lets you do that easily.

I'm sure others can weigh in with a list of the processes and scripts they like.

 

So usually after I stack in DSS I do a few basic things in Photoshop: levels (black point & white point), RGB balance, contrast/brightness, saturation, sharpen, and despeckle. How would these translate into PI functions?



#7 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 26,034
  • Joined: 10 Jan 2014

Posted 12 January 2015 - 07:31 PM

I agree with Goofi here. Focus on a few key tools in PI. I like to think of PI as a collection of indepentent tools or applets, rather than one gigantic, daunting program. That's really what it is, few of the tools depend on each other (mainly the only interconnectivity is with masking.) 

 

There are some key tools you really want to learn, like DynamicBackgroundExtraction, TGVDenoise and MultiscaleMedianTransform|MultiscaleLinearTransform (the latter has been transformative for my work), and the BatchPreprocessing script (which is what does your calibration, registration and integration.) Start with those, experiment with and learn each tool.

 

There are a couple key things to learn and understand about PI as well. First is the UI. It's different, but extremely fluid and powerful once you get used to it. Point to EVERYTHING, and let the cursor hover for a moment. Very informative tooltips appear over just about every text box, check box, slider and other control in a tool dialog. Read those, they will help you immensely. A lot of things are draggable and instancable. For example, you can drag the main tab of a viewport to the background to clone an image. You can drag the little half box (small triangle) icon at the bottom of most tool/script dialogs to the background to save an instance (which includes the current state of the tool). Instancing tools is extremely useful and powerful, as it lets you try out different sets of settings, and quickly switch back and forth between them to see differences in how they affect the image.

 

The biggest thing you want to learn is previews. I didn't find any good tutorials that actually explained WHY you want to use previews...but there are a few key reasons. First, previews are often used by certain tools, like BackgroundNeutralization or ColorCalibration, and they must be drawn around specific areas of an image for those tools. That is probably the first use case for previews you will encounter, and it's a pretty strait forward one. This is not the true power of previews, though.

 

You can also create any number of preview frames on an image, and each preview gets another tab in that viewport's tab list. You can draw previews around small to medium sized areas of an image (well, any size, really, even the whole image...but generally you want them smaller, and around key areas of your image, areas of detail, or example background areas, etc.), then switch to the preview tab, and apply a tool (by dragging that same triangle icon you use to instance) to the preview image. That will apply whatever tool you are using...say MLT or TGV or Curves or whatever...to just the pixels of that preview. Even better, you can make tweaks to the settings of the tool, and reapply again without having to undo. This can DRAMATICALLY increase the speed of your workflow, as you can preview the changes a particular tool will cause to your image without actually having to apply to the whole image. You can also quickly switch back and forth between the original state and modified state in a preview with CTRL-SHIFT-Z.

 

If I had understood the behavior of previews when I first started using PI about 8 months ago, I'd have learned it a lot faster. Use previews! They are the fastest way to learn what each tool does, and what changes to the settings of each tool do to image. There are caveats here. Some tools don't work with small previews, as they are based on the whole contents of the image (i.e. HDRMultiscaleTransform). Other tools are dynamic, and are intended to be applied to and tried based on the whole image. DynamicBackgroundExtraction (DBE) is an example...however DBE allows you to test and preview what it will do in a different way (by default, it does not apply to the associated image...so you can generate a new copy of the image with DBE each time you tweak it and see the results.) 

 

Trial everything. Testing out your settings and seeing how they impact the image is how PI works. You'll find common settings for certain tools by reading tutorials, such as on LightVortexAstronomy. Those are useful starting points, but every camera is different, noise changes with temperature, sub count, etc. To get he best results, you have to try, apply, and try again. Previews will increase the rate at which you can do this 10-100 fold over applying to the full original image then undoing (especially with iterative tools like TGV, where you might need to apply 250-500 iterations, which on big images can take a good long while to complete.) Previews. Your very best PI friend. ;)



#8 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,974
  • Joined: 18 Jul 2011

Posted 12 January 2015 - 07:36 PM

Everyone has their processes and a routine. There are a few things I've found to be critical though. First up, do a dynamic crop after carefully examining the edges all around. This makes such a huge difference in everything else you do. Next up would be either the automatic or dynamic background extraction.

 

Bottom line, working in linear and non-linear make a big difference. PS has no equivalent because the data is stretched to some degree first.



#9 Goofi

Goofi

    Cosmos

  • *****
  • In Memoriam
  • Posts: 8,137
  • Joined: 03 May 2013

Posted 12 January 2015 - 08:36 PM

You're getting good advice from Jon and Terry  and the others  :)

 

Pixinsight lets you do a lot of the processing before stretching the data - think of it as before doing levels and curves in Photoshop. 
This is important because you're keeping the data linear, so you're keeping the relationships between pixel values.

Once you stretch the data, there's no going back.

 

Madratter has a basic workflow on his website - he's got the #2 post here on this thread; in his signature is a link to his website.

You can find his Pixinsight Tutorials there, as well as a basic workflow.

 

After a while, you'll start to develop your own workflow and style - far more so than I did with Photoshop.

To me, Pixinsight is an operating system; each process or script is an application I can run depending on my needs.

It's powerful, and I've only scratched the surface - many here are far more advanced than I am.



#10 D_talley

D_talley

    Fly Me to the Moon

  • *****
  • Posts: 5,247
  • Joined: 07 Jul 2005

Posted 12 January 2015 - 09:26 PM

Pixinsight is a very complicated and involved software package that should not be this hard to learn or use.  I see this question often asking about tutorials on how to use it. I bought the software thinking that I would be able to use it after a short time but I have yet to use it past combining bi-color images. I don't know what I am doing, just following the steps someone tells me to do.

 

Have you ever wanted to learn to dance the tango? You can get numbered cutouts of foot steps and place them on the dance floor and follow them. No, you are not dancing the tango when you follow the steps because you don't know why you are doing what the steps tell you to do. As with the tutorials that several people have made, as nice as they are they are just steps without the reason or background that you really need to create the best image you can.  Each tool, like the DynamicBackgroundExtraction, has a documentation link on the bottom right but the link tells you "Tool documentation not available"?!?!?  You would think that this would be a big help to the new user. So you click on the Resources tab at the top and there are links to their tutorials and videos.  All lack the basics for anyone starting out with the software. 

 

The responses about also show what we as new users have to deal with. Everyone has their own way of doing the processes. I am ok with that but if someone says to use tool XYZ and there are no instructions available, how do I learn how to use it other than hitting a random button to see what happens?

 

My use of PI will be limited until I find a good tutorial that starts from square one.



#11 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,974
  • Joined: 18 Jul 2011

Posted 12 January 2015 - 09:34 PM

Pixinsight is a very complicated and involved software package that should not be this hard to learn or use.  I see this question often asking about tutorials on how to use it. I bought the software thinking that I would be able to use it after a short time but I have yet to use it past combining bi-color images. I don't know what I am doing, just following the steps someone tells me to do.

 

Have you ever wanted to learn to dance the tango? You can get numbered cutouts of foot steps and place them on the dance floor and follow them. No, you are not dancing the tango when you follow the steps because you don't know why you are doing what the steps tell you to do. As with the tutorials that several people have made, as nice as they are they are just steps without the reason or background that you really need to create the best image you can.  Each tool, like the DynamicBackgroundExtraction, has a documentation link on the bottom right but the link tells you "Tool documentation not available"?!?!?  You would think that this would be a big help to the new user. So you click on the Resources tab at the top and there are links to their tutorials and videos.  All lack the basics for anyone starting out with the software. 

 

The responses about also show what we as new users have to deal with. Everyone has their own way of doing the processes. I am ok with that but if someone says to use tool XYZ and there are no instructions available, how do I learn how to use it other than hitting a random button to see what happens?

 

My use of PI will be limited until I find a good tutorial that starts from square one.

 

I'll say this again. If you want to understand how to use the processes, spend up and get the Warren Keller/Rogilio Bernal Andreo tutorials. Also, spend time with the documentation that exists. People right here share a lot on how they use many of the processes. So there isn't a manual with the software....it just takes some effort on our part



#12 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 26,034
  • Joined: 10 Jan 2014

Posted 12 January 2015 - 09:45 PM

Dwight, PixInsight is a complex program, for sure. That's why I say it helps to think of it as a collection of independent apps, rather than one gigantic program with an insurmountable learning curve. Most of the tools are independent of each other, and can be used and learned in isolation. Some of the tools themselves are still somewhat complex, but at least they are more bite-sized. Rogelio Bernal Andreo, an excellent imager, has put together a useful reference guide for PixInsight:

 

http://www.deepskyco...com/PixInsight/

 

This isn't a step-by-step "do it this way", it's just a reference guide. There are some great tutorials that somewhat follow such an approach, such as those on Light Vortex Astronomy, Harry's PixInsight Videos, etc. The thing about PixInsight, and maybe just the nature of astrophotography, is these can only ever be stepping stones. I'm not trying to give people a workflow with what I said about previews before...I simply think that "try and apply" is how PixInsight works. You cannot use one instance of DBE for every single image...for it to be effective, you have to fine-tune it's application to each and every image you process. Why? Because every single image is going to have different background issues. You are unlikely to find a one-size-fits-all noise reduction routine that works magic on every single image. Can you use MLT with the exact same settings for every single image? Sure. Are the results going to be acceptable for every single image? Nope! 

 

Every image is different. You aren't guaranteed to have the same amount of skyfog/lp, the sensor temperatures could be different from night to night and thus dark current noise could change, the sensor orientation to light sources that cause gradients won't be the same for every object, your total integration times will vary, etc. There is no one-size-fits-all "just follow these specific steps with these specific settings" when it comes to getting the best results out of processing your astrophotography. You have to apply the settings that work for each and every image. Being able to do that, with previews on a small subset of pixels from your image, quickly and repetitively until you find the settings that work ideally for each specific image is key to using PixInsight most effectively. You can blast through a dozen different settings for say MLT or TGV, tweaking and fine tuning the settings to give you the best result, in a matter of minutes using previews. You can also see, interactively, what changing each setting of a tool does, and that...at least for me...was far more helpful than any amount of reference guide or "Follow steps 1, 2, 3 to process a galaxy" kind of tutorials. SEEING what PixInsight's tools did with my pixels, quickly and iteratively as I tweaked one slider up and down, with rapid feedback of what those changes did to the data, was my big breakthrough with PixInsight. Before that, I too spent countless hours looking for better documentation, better tutorials, watching every video I could find (sometimes a few times over), and applying the cookie-cutter settings to each and every image and never quite being satisfied with the results. 

 

I never apply the same settings for each tool to every image anymore. I customize every tool's settings to the image at hand. I have my own workflow, what fits my thinking and works for me. I have evolved it a couple of times, but, workflow is workflow, I think that's more personal. Fundamentally, though, I think taking the "try and apply" approach with each tool, regardless of what order you may apply tools in sequence on your images, or whether you apply them to linear or non-linear data, is just how PixInsight works. You can't know ahead of time how one particular set of settings for a tool is going to affect any given integration, whether you need to use a mask with it or not, etc....until you try it. Try and apply! :)



#13 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,324
  • Joined: 11 Aug 2008

Posted 12 January 2015 - 09:49 PM

>>>>>My use of PI will be limited until I find a good tutorial that starts from square one.

 

I think the most comprehensive tutorial is Warren Keller's. It does cover everything, as far as I can tell, and does it in a sequence that helps you get moving pretty quickly, particularly if you take advantage of his process icons.

 

 

http://www.ip4ap.com/

 

is the source

 

They do cost.....but doesn't everything.

Alex



#14 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 26,034
  • Joined: 10 Jan 2014

Posted 12 January 2015 - 10:04 PM

Also, don't underestimate the value of the PixInsight forums. There are many very knowledgeable, helpful members there. 

 

Warren Keller has some free videos. I'd give them a try. I had a hard time with the way he presented (just a little slow...in the rate at which he moved through the program, and even in the way he would say things, terms, interface element names, etc., and I move very fast, so it was just a little frustrating having to wait), and it seemed like most of it was done with a fairly old version of PixInsight (1.7.x)...so I never purchased them. He does cover EVERYTHING, though (that's just based on the free videos he has).



#15 bjgiii

bjgiii

    Viking 1

  • *****
  • Posts: 682
  • Joined: 25 Feb 2014

Posted 13 January 2015 - 06:26 AM

I have found Warren's video very helpful. There's a lot of great information in them. I decided to start with PI instead of PS, so it's all I know when it comes to processing software. I have watched Harry's videos and read and used the workflow from lightvortexastronomy and found both helpful. I'm still very much a beginner with the processing thing, but it is getting easier.



#16 Alex McConahay

Alex McConahay

    Hubble

  • *****
  • Posts: 13,324
  • Joined: 11 Aug 2008

Posted 13 January 2015 - 10:04 AM

Can't you speed up Warren, or just use the closed caption?

 

I find the voice and pace a little soporific, but appreciate it that it helps me take notes as I go.

 

Alex



#17 bjgiii

bjgiii

    Viking 1

  • *****
  • Posts: 682
  • Joined: 25 Feb 2014

Posted 13 January 2015 - 11:43 AM

Can't you speed up Warren, or just use the closed caption?

 

I find the voice and pace a little soporific, but appreciate it that it helps me take notes as I go.

 

Alex

 

He does speak very slow in the videos. I've honestly dosed off watching them once in awhile. :lol:



#18 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 26,034
  • Joined: 10 Jan 2014

Posted 13 January 2015 - 12:12 PM

 

Can't you speed up Warren, or just use the closed caption?

 

I find the voice and pace a little soporific, but appreciate it that it helps me take notes as I go.

 

Alex

 

He does speak very slow in the videos. I've honestly dosed off watching them once in awhile. :lol:

 

 

:lol: That was my problem with his videos. Soporific is an excellent word...but it's almost like it's deliberately soporific, and just deliberate, somewhat syllabic speech in general. Kind of like a kindergarten teacher teaching five year olds. ;P Anyway, I did find that kind of frustrating when trying to learn from the free videos he has, it's tough to stick with it throughout a whole video...you either get bored, or your just in "Come on, Come on, Get movin on!" mode the whole time. Each video could be half as long if he would pick up the pace a bit. People can always rewind and replay if they need to hear something more thoroughly. I guess you could play the videos back at a faster rate, and tweak the audio pitch to get through them faster...I never tried that. :p



#19 Peter in Reno

Peter in Reno

    Hubble

  • *****
  • In Memoriam
  • Posts: 13,067
  • Joined: 15 Jul 2008

Posted 13 January 2015 - 12:19 PM

If you don't like Warrne's voice, disable the sound and enable Closed Caption. Besides I can't hear his voice because I am deaf. ;)

 

Peter



#20 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 26,034
  • Joined: 10 Jan 2014

Posted 13 January 2015 - 01:24 PM

Does he have closed captions? If he does, I might just do that. There are some videos that cover things in PI I haven't really explored yet. 



#21 Peter in Reno

Peter in Reno

    Hubble

  • *****
  • In Memoriam
  • Posts: 13,067
  • Joined: 15 Jul 2008

Posted 13 January 2015 - 01:33 PM

Warren Keller's PI Part 1 and 2 video tutorials have built-in CC. Not sure of Part 3 because I didn't buy it.

 

Peter



#22 mostlyemptyspace

mostlyemptyspace

    Messenger

  • *****
  • topic starter
  • Posts: 444
  • Joined: 05 Jan 2014

Posted 13 January 2015 - 02:14 PM

Harry has some excellent video tutorials. They are what got me started.

You can try my PixInsight tutorial that takes you through processing some Messier 33 data. That is on my website.

 

Madratter your tutorial is awesome. I read through it once and I'm going to go through it again now and follow along in PI. It would be amazing if you could add a section on the initial stacking though. That's a big component to the workflow that you skipped. 



#23 Madratter

Madratter

    Hubble

  • *****
  • Posts: 13,277
  • Joined: 14 Jan 2013

Posted 13 January 2015 - 02:30 PM

Thanks! I skipped the stacking because people have different preferences such as Deep Sky Stacker and even in PI there are multiple methods. The easiest in PI and where I would suggest starting is using the Batch Preprocessing Script. It would be a good idea for me to add that at some point.

#24 Goofi

Goofi

    Cosmos

  • *****
  • In Memoriam
  • Posts: 8,137
  • Joined: 03 May 2013

Posted 13 January 2015 - 03:04 PM

I prefer to read a page than watch a video; I much prefer lightvortex and Madratter's tutorials to a video. But, I do have Warren's videos, and while they're slow (annoyingly so), I have gotten a lot of value out of them.

 

Right now, this guy's web pages are where I'm spending most of my time:

http://www.radice.bi...li/translations

What he does with pixel math is impressive ...



#25 mostlyemptyspace

mostlyemptyspace

    Messenger

  • *****
  • topic starter
  • Posts: 444
  • Joined: 05 Jan 2014

Posted 13 January 2015 - 11:01 PM

Does anyone have some good raw data from a DSLR I can use to practice? It seems most of the data I find are either in the final stacked TIF/FIT, or spit into LRGB. Since I have a DSLR, I don't have separate LRGB channels. I don't know when I'll have a chance to get some good data of my own, so I'd like to practice.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics