#1  
Old 21-11-2018, 11:40 AM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 948
A pixinsight processing question

What is the best order for processing the image while still linear?
To start I crop, remove background gradients (with DBE) and then background neutralise & colour calibrated.

After this, should I do noise reduction before, or after sharpening with deconvolution and/or multiscalelineartransform. I can see the point of doing it in either order so am confused.

NB -I also do some noise reduction on the image after stretching.
Reply With Quote
  #2  
Old 21-11-2018, 03:44 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,579
Quote:
Originally Posted by ChrisV View Post
What is the best order for processing the image while still linear?
To start I crop, remove background gradients (with DBE) and then background neutralise & colour calibrated.

After this, should I do noise reduction before, or after sharpening with deconvolution and/or multiscalelineartransform. I can see the point of doing it in either order so am confused.

NB -I also do some noise reduction on the image after stretching.
The less processing done before decon the better, so do it before noise reduction. Don't do any other type of sharpening until the end of your workflow.

Doing additional noise reduction later is fine. I try to err on the side of caution and only do a very light noise reduction while the data is linear.

Cheers,
Rick.
Reply With Quote
  #3  
Old 21-11-2018, 07:30 PM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 948
Thanks Rick. That clears up several questions
Reply With Quote
  #4  
Old 23-11-2018, 01:13 PM
sil's Avatar
sil (Steve)
Not even a speck of dust

sil is offline
 
Join Date: Jun 2012
Location: Canberra
Posts: 1,203
Chris, my workflow and i guess most others, evolves over time as i learn new methods andget to understand what each tool brings to the workflow.

To that end I have moved to using the desktop of pixinsight to build my workflow and evolve it by dragging each process to the desktop so its config is ready to rock on the next set of data later on. So I put my first icon at the top and put more underneath then start a new column until its all there. Plenty of space to split off at any point and put two side by side with different settings then work through the workflow twice to compare how the outcome is effected by that one change. Its helpful too when you arent sure about the values you set to then run another dataset through and see if the settings are too aggressive or not. Typically we capture with the same gear so the workflow should work equally well on all data, until it comes to masks where the shot itself determines the mask/change.

Along side this I have a note system too to fill out detail about each step, as somtimes there are things to look out for in particular that you want to be away of at a particular step. Generally each step is either enhancing Signal or reducing noise. Every step usually acts on both, one to your benefit and the other to the detriment of the data. Each step can compound what happened in previous steps so maybe you over do something early on that introduces a little noise you cant notice but by the time your workflow is done that noise is greatly amplified.
By workflow I mean from subs set through to the jpeg you post on facebook. So your final image may look better if you do noise reduce early or late depending on your use of the final images.
I tend to prefer doing lighter adjustments (less aggressive) at various stages rather than one big processing step at one point. But maybe I'll change my mind later on about that.
Keep in mind every time you make the image look better you are probably going to have extra noise to deal with later. Pixinsight has several tools that are really duplicates of each other just configured for different uses in the workflow, and i think ultimately everything can be done purely with pixelmath. So I go through others tutorials and pick up when and how people use these tools and why so i can try alternative approaches in my own. Its a perpetual tweaking machine And then a PI update adds something that negates a large chunk of your workflow with a single tool.
Reply With Quote
  #5  
Old 23-11-2018, 11:32 PM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 948
Thanks sil. Thanks kind of what I do. Except I have a bunch of different versions of desktops/projects that I've cobbled together from different sources. As you say it constantly evolves. I suppose it's silly to think that one workflow will suit all images.

I like your suggestion of having a fork in the road. Then compare. I'll try that!
Reply With Quote
  #6  
Old 24-11-2018, 09:03 AM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,579
I think it's a lot like cooking. You start by following a recipe exactly. As you gain experience you borrow bits from other recipes and start to improvise a little. Eventually, you know instinctively what to throw in the bowl and just season to taste as you go
Reply With Quote
  #7  
Old 24-11-2018, 12:27 PM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 948
Thanks Rick So I won't do what my son does and drown everything in tomato sauce
Reply With Quote
  #8  
Old 24-11-2018, 12:48 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,579
Quote:
Originally Posted by ChrisV View Post
Thanks Rick So I won't do what my son does and drown everything in tomato sauce
That's a phase we all go through, Chris
Reply With Quote
  #9  
Old 26-11-2018, 12:01 PM
sil's Avatar
sil (Steve)
Not even a speck of dust

sil is offline
 
Join Date: Jun 2012
Location: Canberra
Posts: 1,203
Quote:
Originally Posted by RickS View Post
I think it's a lot like cooking. You start by following a recipe exactly. As you gain experience you borrow bits from other recipes and start to improvise a little. Eventually, you know instinctively what to throw in the bowl and just season to taste as you go
And now I crave spag bol , thanks ;p
lol
Reply With Quote
  #10  
Old 26-11-2018, 12:48 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,579
Quote:
Originally Posted by sil View Post
And now I crave spag bol , thanks ;p
lol
Just don't apply too much deconvolution or the spaghetti will look wormy
Reply With Quote
  #11  
Old 26-11-2018, 04:17 PM
ChrisV's Avatar
ChrisV (Chris)
Registered User

ChrisV is offline
 
Join Date: Aug 2015
Location: Sydney
Posts: 948
I find I can't use very much sauce with deconvolution compared to what I've seen in various guides (lightvortexastronomy, John Rista and the Inside IP book by Keller).

I'm using a range-star mask to just work on the target plus a star mask for local support - but usually I can only go up to about 10-15 iterations with really low global dark for deringing (< 0.001). Any more than that and the worms appears.
Reply With Quote
  #12  
Old 26-11-2018, 06:27 PM
RickS's Avatar
RickS (Rick)
PI cult recruiter

RickS is offline
 
Join Date: Apr 2010
Location: Brisbane
Posts: 10,579
Quote:
Originally Posted by ChrisV View Post
I find I can't use very much sauce with deconvolution compared to what I've seen in various guides (lightvortexastronomy, John Rista and the Inside IP book by Keller).

I'm using a range-star mask to just work on the target plus a star mask for local support - but usually I can only go up to about 10-15 iterations with really low global dark for deringing (< 0.001). Any more than that and the worms appears.
I have found the occasional data set that can take a lot more but I often use only a handful of iterations.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 03:22 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
SkyWatcher WiFi Adaptor
Advertisement
SkyWatcher 2018 Catalogue
Advertisement
NexDome Observatories
Advertisement
Lunatico Astronomical
Advertisement
OzScopes Authorised Dealer
Advertisement
Interest Free Finance
Advertisement
Astronomy and Electronics Centre
Advertisement