patchy background at the end of my process

Frédéric Ruciak

Hi all,

I am trying to process my third image combining two capture sessions (M51 120x60s +180x30s)) with dedicated bias, darks, flats. WBPP (with keywords grouping) worked fine with cosmetic correction. I had to crop because the two sessions were not framed exactly the same (oops). I had to remove some gradient with DBE then SPCC, BXT, NXT and here comes my problem. I am trying to get both a luminous galaxy with some (but not too much) colors, without a burned core. So I tried using MaskedStretch. It works well but at the end I always get some patchy background with some red flavour somewhere, it's not really strong but really not good. I clearly noticed that if I want to push maskedstretch too much (200 iterations) with an increase of the stretched background image the patchy background appears (with reddish areas) more and more. I tried to interleave some HT to darken the background a little bit before applying the maskedstretch but if I go too far maskedstretch sends an error message (too dark image). I have to add that I tried to play a little bit with color saturation and curves to bring some colour in the galaxy at the linear stage.
The final result is not a disaster but I am willing to get the best of what I can do from my data.

Is there something wrong in my workflow or what should I take care of to avoid this emergence of the background at the end? Thanks in advance, Frederic:)" style="box-sizing: border-box; border-style: none; max-width: none; height: 22px; vertical-align: text-bottom; width: 22px; background: url("styles/default/xenforo/smilies/emojione/sprite_sheet_emojione.png") 0px 0px / 100% no-repeat rgb(254, 254, 254); color: rgb(20, 20, 20); font-family: Nunito, "Helvetica Neue", Helvetica, Roboto, Oxygen, Ubuntu, Cantarell, "Fira Sans", "Droid Sans", sans-serif; font-size: 15px;" />

BlockHead

HI Frederic,

This is an example where I think practice needs to play a part of the answer. 

The first thing that strikes me is that the galaxy is green. (It shouldn't hurt anything... but why did you run SPCC twice?). 

1. What did the image look like after the first instance of SPCC? It looks like you have some good data... 

2. You did quite a bit of stretching... I think you may need to think about choosing stretches that only move you in incremental steps "forward" ... you pretty much never want to apply a process that isn't 100% in a "positive" direction. (Sometimes people make adjustements... and then make another adjustment to fix their last one... and so on). 

The patchy background is hard to see in the JPEGs you posted (but I believe you). Since the color balance of the bright signal (galaxy) is off... it might influence your sky color as well.

So my recommendation (a general suggestion) is to go back in time and focus on making only absolutely improvement choices. If a process does not improve the view *ever* time- you must pause and find a better way. 

P.S. 200 iterations of masked stretch does not "push it more" or make a greater effect. Instead it is just taking smaller steps in its calculations. 

Here is my finally note...I would like to start a series of videos in Horizons where I take someone's data... and process it with my choices...  As I go along...I would compare the user's results with my own for instructional purposes. 

If this interests you (if you are willing)- It seems to me this could be an excellent example.

-the Blockhead

Frédéric Ruciak

Hi Adam, thank you for your answer. I like the way that you try to teach how to fish rather than giving some fishes to eat ;-).

I understand that mastering the process is a combination of some guidelines and principles that need to be adapted on each case. Sometimes I lost north and as you say I have to step backward and ask myself why didn't I get what I expected?

I have to say that I am still in the process of digesting fundamentals and the two issues I am really struggling to better understand are colours (clearly something wrong in my attempt to push blue and red in the galaxy) and stretch to get the expected result - despite the remarquable stretch academy :-).

I know that I have still some issues to fix on my acquisition setup too.

But I'll remember your proposal and will be volunteer in the future. Once I will feel ready I'll dot it.

Frédéric

Frédéric Ruciak

Hi Adam,

You mentioned that M51 looks green.

I looked at the video when you present the workflow of M51 (not completely yet) treating an image with some gradient and running SPCC first. I have some gradient including color gradient in my masterlight and DBE was applied before SPCC so I ran an interesting test that I am trying to diagnose.

I took the same M51 light master and applied the same steps twice but with an inversion between SPCC and DBE from the same start: for one workflow DBE SPCC BXT NXT for the other SPCC DBE BXT NXT.  

At the end I am looking at the views with STF linked channels still in linear stage. Of course the DBE process applied on each flow is the same.

The path with DBE first first gets a greenish galaxy not the other!

I would have guessed that SPCC ran after DBE would lead to a cleaner white balance knowing that the white balance is done and kept unchanged from another later process so I am puzzled by the result.

Clearly the fit of SPCC is better when ran first (both sigma are very low) and the values are different.

I attach both jpegs and the pdf of the SPCCs results in the same order

first image DBE first greenish M51

second image SPCC first neutral M51 (as i see it).

I checked on another M51 master from another acquisition with much less background color gradient and the difference is much less perceptible but still there. By the way the scale of the linear image are different as I have to tune STF in a different way to get the same visual result on the screen (photography tuned)

I do not understand why but at least I will NEVER run DBE before SPCC in the future.

Frédéric

BlockHead

I suspect that this has less to do with the order of operations and more to do with your choice/setup of the BN correction in SPCC. 

-the Blockhead