Jump to content


Photo

Learning PixInsight

  • Please log in to reply
295 replies to this topic

#1 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5402
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 17 March 2013 - 06:49 AM

Besides Harry's video tutorials..., are there any decent step-by-step instructions that someone completely new to PixInsight could use from start to finish for astro image processing? Thx

#2 terry59

terry59

    Fly Me to the Moon

  • *****
  • Posts: 5069
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 17 March 2013 - 06:55 AM

http://www.ip4ap.com/

#3 bilgebay

bilgebay

    Skylab

  • *****
  • Moderators
  • Posts: 4254
  • Joined: 06 Nov 2008
  • Loc: Turkiye - Istanbul and Marmaris

Posted 17 March 2013 - 06:56 AM

Yes, have a look at IP4AP website here.

It's not free though.

#4 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5402
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 17 March 2013 - 07:04 AM

Thanks Terry/Sedat; I was looking for something like I... and Scott... have produced, free/practical online instructions. Thx

#5 terry59

terry59

    Fly Me to the Moon

  • *****
  • Posts: 5069
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 17 March 2013 - 07:08 AM

There are a number of those on the PI site

#6 hytham

hytham

    Apollo

  • *****
  • Posts: 1323
  • Joined: 25 Dec 2012
  • Loc: Northern Tundra

Posted 17 March 2013 - 03:42 PM

Rogelio Bernal Andreo aka my freaking inspiration.

http://www.deepskyco.../tutorials.html

#7 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1906
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 17 March 2013 - 04:45 PM

Harry's tutorials are the best. Add to that a basic workflow as follows and you are on your way

Stack (calibrate+integrate) -> crop boundaries
Gradient removal and color correction (DBE, Color calibration, SCNR)
Noise removal via Atrous Wavelets
---
Stretch - Histogram transfer (levels)
HDR wavelets to increase dynamic range
LHE Local contrast enhancement
ACDNR, Nose reduction
Saturation Boost via curves

The dashed line separates the first three steps you do on linear data. The latter are done on stretched and therefore non-linear data. Harry has a tutorial for each of these major steps. I do these steps (in addition to others) on every single image I process.

Ajay

#8 srosenfraz

srosenfraz

    Skylab

  • *****
  • Posts: 4138
  • Joined: 06 Mar 2011
  • Loc: United States

Posted 17 March 2013 - 07:32 PM

Rogelio Bernal Andreo aka my freaking inspiration.


Ditto.

#9 kbev

kbev

    Vostok 1

  • -----
  • Posts: 183
  • Joined: 29 Dec 2010
  • Loc: Far, far east Mesa

Posted 18 March 2013 - 01:31 AM

There are a number of those on the PI site

+1, in particular I have this one bookmarked so I can refer to it as needed when processing images: Processing a M45 DSLR image There are others dealing with using specific PI processing tools but as a general overall flow I like referring to this one to help me learn to navigate PI.

#10 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5402
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 18 March 2013 - 02:30 AM

Thanks hytham/Ajay/Scott/Kevin.

Ajay, your outline is very helpful for someone completely new to this; exactly what I was looking for at this time.

Everyone, this is just a start; will have more specific questions as I try tackling PixInsight. Keep the discussion going, keep your feedback coming, and let's get more granular as we more forward. If possible and if all goes well, I would like to come up with an instructions doc with your help. Regards

#11 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1906
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 18 March 2013 - 09:43 AM

You are welcome Mike. I myself started with a basic outline from another imager I know (Rainycityastro on CN). One thing I left out is Morphological Transform to reduce stars.

Most of my DSLR images were captured with Nebulosity and I saved the images as compressed FITS. Pixinsight's stacking routine has problems with this. So with these images I do calibration and stacking with the new batch process facility in Neb and then move to PI for registration and final stacking.

PI can handle these files if you first convert them, but that takes a while.

Rogelio's site has some great bits of information too, but Harry teaches you the nuts and bolts of various controls in PI with examples.

Ajay

#12 guyroch

guyroch

    Vendor (BackyardEOS)

  • *****
  • Vendors
  • Posts: 2955
  • Joined: 22 Jan 2008
  • Loc: Under the clouds!

Posted 18 March 2013 - 09:52 AM

I just bought Part 1 and Part 2 from IP4AP .

I really need all the help I can get for PixInsight... powerfull... but steep learning curve.

Guylain

#13 hytham

hytham

    Apollo

  • *****
  • Posts: 1323
  • Joined: 25 Dec 2012
  • Loc: Northern Tundra

Posted 18 March 2013 - 09:29 PM

I just bought Part 1 and Part 2 from IP4AP .

I really need all the help I can get for PixInsight... powerfull... but steep learning curve.

Guylain


Just keep on playing with it and you will come to discover that all of your thoughts about its difficulties were a slight exaggeration. I started learning both PS and PI at the same time and quite honestly I prefer the work flows in PI and I absolutely love the scripting aspect of it - so much control. It's fantastic.

You'll never look back :)

If you have not already signed up on the PI Forums, do so immediately. Wealth of information!

#14 harry page 1

harry page 1

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 25 Jul 2009

Posted 19 March 2013 - 06:55 AM

Hi
what can I do to help make the basics better to understand

Always open to suggestions

Harry

#15 jsines

jsines

    Mariner 2

  • *****
  • Posts: 256
  • Joined: 06 Sep 2011
  • Loc: Berkley. Michigan

Posted 19 March 2013 - 12:49 PM

Hi
what can I do to help make the basics better to understand
Always open to suggestions
Harry


Hi Harry - thanks for popping up here, I really appreciate your videos/tutorials. They've helped me come a long way in my processing skills. I bought PI about 6 months ago, but I still feel like a newbie.

One item I'm noticing in the tutorials and the videos is the focus on CCD cameras over DSLR cameras. I'm using a modded Canon 450D, I'm shooting inside a white zone, and I'm having a lot of issues dealing with noise. I feel I've got DBE and ABE down, but noise keeps frustrating me. I've read all the threads on the PI forum about noise, but I always seem to overdo it on the noise reduction and end up with lumps.

Not sure if you have the equipment or if someone else can point me to the tutorial, but I'd like to see more examples of processing with DSLRs in bad conditions. The data in the tutorials I've watched/read is so clean to begin with compared to my DSLR exposures, it looks like Hubble data.

(I know I also have my own issues about expecting too much data in the little imaging time I get, but I'm working on that. I must repeat - I can't get 6 hours of data out of 1 hour of imaging time. :p )

Thanks!

#16 harry page 1

harry page 1

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 25 Jul 2009

Posted 19 March 2013 - 01:54 PM

Hi
I can assure you that I shoot my images from a very orange place:)
we all have to learn to get the best from our images and data , where ever taken .

Could you let me have some of your data and let me look at it and see if I can help
Harry

#17 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5402
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 19 March 2013 - 02:05 PM

what can I do to help make the basics better to understand


Add to that a basic workflow as follows and you are on your way

Stack (calibrate + integrate) -> crop boundaries
Gradient removal and color correction (DBE, Color calibration, SCNR)
Noise removal via Atrous Wavelets
---
Stretch - Histogram transfer (levels)
HDR wavelets to increase dynamic range
LHE Local contrast enhancement
ACDNR, Nose reduction
Saturation Boost via curves

The dashed line separates the first three steps you do on linear data. The latter are done on stretched and therefore non-linear data. Harry has a tutorial for each of these major steps. I do these steps (in addition to others) on every single image I process.



Harry, thanks for offering to help; greatly appreciated!


For the start, above is a high-level outline by Ajay if you could validate which I am pretty sure would be agreed upon succession and totality of basic processing steps. Once we have this high-level workflow laid down, then we can get into the details and granularity of each in methodical manner. Look forward to your validation and/or suggestions? Regards

#18 harry page 1

harry page 1

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 25 Jul 2009

Posted 19 March 2013 - 02:40 PM

Hi

Yes its about right , but of cousre there may not be a need for hdr wavlets or lhe , depends on your image :cool:

And of course sometimes less noise reduction is better than too much

Harry

#19 jsines

jsines

    Mariner 2

  • *****
  • Posts: 256
  • Joined: 06 Sep 2011
  • Loc: Berkley. Michigan

Posted 19 March 2013 - 02:51 PM

Thanks for the offer to help, Harry. The data is on the Rosette Nebula, I posted an inquiry thread in the Image Processing Challenges subforum over at the PixInsight forum explaining my problem. I posted all the lights, the master dark, the master bias, and the master flats on the pixinsight server under the jsines folder. I'm in between light pollution filters because I went from a 1 1/2 inch adapter to a 2 inch adapter to get rid of vignetting. The Astronomik CLS should arrive tomorrow.

I don't mean to cross streams here by referencing another forum, not sure if it's allowed. :p

From the pixinsight post -

* Orion ED80T
* Orion Sirius mount, unguided
* Canon 450D, modded, using a 2 inch adapter
* 50 x 120 second lights at 400 ISO (1 hour, 40 minutes) over 3/3/13 and 3/4/13
* 20 darks at 400 ISO
* 20 flats at 100 ISO
* 100 bias
* stacked and processed in PixInsight
* imaged in a white zone

This is probably processing attempt #6, and I'm still not completely happy with it. I don't know what I can do or if I can do anything. I started imaging at the meridian and imaged until it was about 25 degrees from the horizon. I'm at 42 degrees North in the US. I'm in a white zone, so I'm wondering if that's an issue.

* stacked using the BatchPreProcessing script - I ended up with average SNR increases of 14, 16, and 13 (RGB)
* integrated using windsorized sigma clipping
* cropped
* used Dynamic Alignment to match a jpeg image from the APOD website to the cropped image as a guide for DBE. I probably put about 10-15 DBE samples in the whole image.
* then processed as I usually do - AtrousWavelets, Background Neutralization, Color Calibration, Histogram, SCNR, ACDNR, HDRWavelets, Histogram, Curves, Histogram, LHE, etc.

I couldn't exactly get the red like I wanted, it's probably over-saturated, and I see I need a field flattener. I'm also wondering if it's a combination of using a dslr and imaging in a white zone that's preventing me from getting a better image. I'm also totally open to the idea that I messed up somewhere. I've tried this object like 3 times over the past few months, and I've ended up throwing it all out because I'm not getting anything good enough to show anyone.

Attached is the best I've been able to come up with so far, and I'm not sure how to proceed - I'm totally open to feedback.

Attached Files



#20 Alfredo Beltran

Alfredo Beltran

    Viking 1

  • -----
  • Posts: 500
  • Joined: 01 Aug 2010
  • Loc: Bogota, Colombia

Posted 19 March 2013 - 04:25 PM

I've used this tutorial with good success. It's in Spanish but you can follow it. Has a lot of useful information.

Best regards

Alfredo

#21 harry page 1

harry page 1

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 25 Jul 2009

Posted 19 March 2013 - 05:21 PM

Hi
Ok I will download the files and have a play

Give me a day or so and I will get back to you

Harry

#22 jsines

jsines

    Mariner 2

  • *****
  • Posts: 256
  • Joined: 06 Sep 2011
  • Loc: Berkley. Michigan

Posted 19 March 2013 - 06:49 PM

Thanks! I'm also redoing them after learning a few things in the past week. I also didn't have the DSLR_RAW settings changed in Format Explorer. I changed them, but I can't find on the forums if that would have affected the output.

#23 mmalik

mmalik

    Fly Me to the Moon

  • *****
  • Posts: 5402
  • Joined: 13 Jan 2012
  • Loc: USA

Posted 19 March 2013 - 06:56 PM

Asking this for learning sake...

How does DBE Magic compare to GradientXTerminator; are these doing necessarily the same thing or one doing something different/additional than the other? If one had both, would running both be an over-kill/moot? If one had both and were to pick one, which one would be preferred? Thx

#24 Peter in Reno

Peter in Reno

    Fly Me to the Moon

  • *****
  • Posts: 6076
  • Joined: 15 Jul 2008
  • Loc: Reno, NV

Posted 19 March 2013 - 07:00 PM

I feel that PI DBE works better than GradX. DBE appears to be easier and more powerful.

Peter

#25 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1906
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 19 March 2013 - 07:26 PM

Just to be clear, light pollution gradients are not noise. These are removed by the DBE or ABE tools.

All the images I have posted in this form were acquired with a Canon 350D which is quite a bit noisier than the newer DSLRS most people are shooting with these days. I do an Atrous Wave subtraction as demonstrated by Harry and some times I do a *mild* GreyCStoration correction directly afterwards. This is done on unstretched data. After stretching I do ACDNR for a little more noise reduction.

With this strategy I could retrieve a decent image at the top of this post. To see how noisy the stacked image was before all these routines, scroll down to the bottom of the same post.

Imaging with a DSLR is going to be inherently noisy. Some of Harry's images, I believe were taken with a CCD with a Sony chip.

I think we should probably get back to the question that started this thread.

Ajay






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics