is my data good enough

[iwols]

hi guys just set my pier up and got the following images, before i carry on capturing i would like to know if its my data or processing skills that are not up to scratch if anyone fancies a quick go at my data i would be thankful heres my poor effort and my original 5 min images lrgb,its just if my data is good enough i can carry on capturing while i learn the processing side better

 

https://www.dropbox.com/sh/h64891fecglg0oi/AADoYouZlhp7h_QCwxc8tn9wa?dl=0

and my feeble effort

[mikey2000]

32 minutes ago, iwols said:

 

 

https://www.dropbox....Cwxc8tn9wa?dl=0

 

I can't get the link to work.  Can you repost it please.  I'm still learning Pix Insight and it's been 100% cloudy since forever here, so any new data would be most welcome :-)

[iwols]

hi mikey sorry try this

https://www.dropbox.com/sh/h64891fecglg0oi/AADoYouZlhp7h_QCwxc8tn9wa?dl=0

[ChrisEll]

Here's my attempt. Your data looks fine. Green sat/plane track is present which should have been rejected during stack.

 

[iwols]

thanks chris really appreciate that,it means i can carry on collecting data(when the clear skies make an appearance) even though im not up to scratch with the processing,youve done a superb job with the 10 mins each of lrgb,do you mind me asking what software you used and the rough workflow ,and again thanks

[ChrisEll]

Thanks iwols. I used PI and GIMP, so the usual PI workflow of LRGB combine, ABE, masked stretch, histo. Then GIMP to finish off with levels and contrast, etc.

I forgot to star align your images which is why the stars look a bit stretched left to right... oops!

[ChrisEll]

Your data capture is very good so keep that up as you get to grips with the image processing side, good luck!

[iwols]

actually purchased a pixinsight licence a while ago but really struggling with it,albeit not had a lot of time with it either ,watched a lot of videos but they all seem to do it different and use flats,bias etc ,if i only knew the basics thanks though chris a massive insentive to keep capturing the data

[iwols]

just a thought chris where in your workflow does star align come in

[ChrisEll]

I don't usually star align manually cos I stack my lights using BatchPreProcessing and that takes care of it but with your data star align should have been my first step. I used PixelMath to average the data for each channel and then LRGB combined the results.

There's a book called Inside Pixinsight by Keller which is a useful read.

[iwols]

thanks chris actually bought that book a while ago 

[iwols]

 

6 hours ago, ChrisEll said:

I don't usually star align manually cos I stack my lights using BatchPreProcessing and that takes care of it but with your data star align should have been my first step. I used PixelMath to average the data for each channel and then LRGB combined the results.

There's a book called Inside Pixinsight by Keller which is a useful read.

going to get the book out and take it one step at a time ,so chris(pretty please) watch this space   before you do the lrgb combine , do you stack the lrgb data individually so you finish with a master lu,red,green and blue

[david_taurus83]

Check out Light Vortex Astronomy. There are some really good step by step tutorials on there for Pixinsight, in particular the pre processing one. I followed that tutorial to have a second go at M81 as I followed a quick you tube how-to first, and the difference is very good. One hell of a learning curve though!

[david_taurus83]

One thing I should add to that though if you do decide to follow the pre processing tutorial, is to perhaps number the folders as you create them after following each step. I started on some data the other night, got so far and had to stop as it was getting late and now I can't remember which step I took last! So need to start again and depending on your pc/laptops processing power, it takes some time.

 

Ps. I'd be happy if I could could gather that data as well as you have. ?

[iwols]

the thing i struggle with is that all the tutorials seem to include dark,flats,bias which i dont use at the minute

[david_taurus83]

5 minutes ago, iwols said:

the thing i struggle with is that all the tutorials seem to include dark,flats,bias which i dont use at the minute

Don't think you can escape those tbh. Darks are open to debate it seems but I've been reading that flats and bias are a must. I'm a novice myself though, no expert. What equipment are you using?  Apologies if all your gear is in your signature, I'm on my phone and it doesn't display signatures.

[mikey2000]

If you don't have Darks, Flats, Bias files etc, you can just skip to the registration/alignment and then stacking steps.

 

You can also throw "dark flats" into the mix to further confuse the issue :-)       With a cooled sensor, you can make a library of these reference frames and re-use them over and over again.  I think maybe it's best to reshoot flat files every now and then (to better remove new dust marks on the sensor/optics) but I haven't bothered yet.

 

I took a look at the data and it looks good to me.    All you need is more of them.  I also started with the Light Vortex tutorials.  It all seems impossibly complicated at first but it starts to get easier with practice.  

 

Here is a my set of steps that I follow (although I still need to refer to the tutorials to get the right settings in each tool....)

calibrate (darks, flats etc)
CosmeticCorrection
SubFrame Selector Script (as described in the tutorials)
registration
LocalNormalisation
Integration
 

Repeat for all 3 R G B and L

Crop
DynamicBackgroundExtraction on all stacks
RGB Combine
PhotometricColourCalibrate RGB

NoiseReduction (Usually MultiscaleMedianTransform)
MaskedStretch then maybe HistogramTransformation

Combine RGB and Lum
more NR?
Colour Tweak?
HDR/LocalHistogram

Save!

 

good luck...  There are plenty of very knowledgeable PixInsight people around here.  I've been given lots of help and great advice.

 

[iwols]

thanks mikey for the workflow

[kens]

A couple of suggestions for your data. First is that you could probably afford to reduce the exposure time on each sub. Possibly 2 mins for RGB, maybe even less. For Lum you could go way shorter. By my calculations you should be aiming for a sky background around the 700 mark, maybe 800 to be safe since i don't know the offset for the camera. Your least  exposed RGB subs (blue as it turns out) are exposed to over 4000. Your Lum subs are over 12000 (as you'd expect). 

Its not as issue as far as saturation goes but you are exposing longer than you need to bury the read noise (which is quite low on this camera). So you could expose for 2 minutes on RGB but take 3x as many subs to get the noise reduction benefit from stacking. Shorter subs also suffer less from satellite trails and the like. On Lum you could go even lower, arguably 30s.

My calculation: 4e- read noise (RN), 0.24e-/ADU gain, 20e- offset (guessed). So for 10xRN squared signal your sky background (histogram peak) should be : (10x(4x4)+20)/0.28 =643 ADU

Based on the data I'm guessing you are in heavily light polluted area? If so you might want to stick with straight RGB imaging. I've found little or no benefit from LRGB at my location. I think maybe the Lum brings out the detail in the LP as much as it does for the target  

[ChrisEll]

13 hours ago, iwols said:

going to get the book out and take it one step at a time ,so chris(pretty please) watch this space   before you do the lrgb combine , do you stack the lrgb data individually so you finish with a master lu,red,green and blue

Yes you need to stack 4 times to produce your LRGB masters. In the book there's a mono workflow on page 247... I would skip Deconvolution for now. Interesting how the book suggests quite a few steps to do before LRGB combine which I would normally do after. Use that as a starting point and let your own workflow evolve from it.

[iwols]

 well i think its an improvement?

just on my first attempt anyway?

[ChrisEll]

Good progress!

[ChrisEll]

Looks grayscale though...

[iwols]

11 hours ago, ChrisEll said:

Looks grayscale though...

yes it does your right,heres my integrated seperate lrgb files i get before rgb comb are they ok? 

https://www.dropbox.com/sh/3y2byl912s7wp2m/AACUyPrP2J_uHe4d4IF9Ydu0a?dl=0

[ChrisEll]

They look fine. I LRGB combined your files and got a very green image when I use STF to view it. I then used HistogramTransformation to correct the G channel (click on the G button and use sliders to move it in line with R&B... use the preview button to see your changes in realtime). After that I used HistogramTransformation again to give it a good stretch. Finally I used ColorSaturation and got this image:

Beyond that some Curves, UnsharpMask and noise reduction will help.