Calibration with Ma...
 
Share:
Notifications
Clear all

Calibration with MasterFlats do not seem to take?  

  RSS

(@jjones)
Brown Dwarf Customer
Joined: 8 months ago
Posts: 7
September 27, 2021 02:11  

I have been getting a little frustrated to say the least... I can't tell what I'm doing wrong as I have the following occurring with my Flats, which seem worse when I add them to my image and nothing seems to get rid of this Oiii Flat issue.

Explanation of equipment:

Rasa 8 with 1600mm monochrome and using Baader f/2 3.5nm/4nm new filters in a bortle 3-4 with some light pollution but not that much.  Oiii displays this very solid artifact when I take flats, not in Sii or Ha, but I figured it should calibrate out with Flats?  since it's in the Flat?

I am using gain 50, and have the following:

Master bias taken with 40 frames, lens covered, and .001s exposures.
Master dark taken with 20 frames, in this case gain 50, exposure was 300seconds, lens covered, same temp, etc.
Master Flats taken with 20 frames, .06s got me to about 30% ADU (shows the artifact as in this image of my flat.  I stretched this 10% in APP to view)

 

APP Flat Gain50 stretched 10

Example of my master flat, stretched in app 10% to show you the artifact... obvious in bottom right.

Stacking in APP, I add my lights, then add my bad pixel map, my master flat, master dark master bias and integrate it... It assignes the masters to all channels and all such, shows it's using MB-1, MF-1, MD-1 and BPM-1.

 

Final Stacks:

APP LBN325 Oiii 50images stretched10 withoutflats

This first image is with everything, except I removed the Master Flat, so I could see if it's even applying the flat.  It actually looks better than when I apply the master flat.

APP LBN325 Oiii 50images stretched10 withflats

This is when I have all my calibrations, including master flat, it gets brighter, doesn't remove what I thought it should, and the artifact in the corner even gets worse...

 

Am I wrong, shouldn't the Master Flat take care of that?

 

Here is my APP and frames and calibrations loaded, seems it's all set to apply?  Any suggestions is greatly helpful... I have rebuilt master flats several times, ran APP now in about 20 different configurations, can't seem to figure it out.

image

ReplyQuote
(@vincent-mod)
Universe Admin
Joined: 4 years ago
Posts: 3961
September 27, 2021 16:59  

So, a flat is used to correct for vignetting in your optics. If the flat is taken correctly (which can be a challenge) then it corrects for the possible vignetting in the setup. Apart from that it gets rid of dust spots. It does not remove gradients that are caused by light pollution. For this the light pollution correction tool is used in the tools-tab. It does happen quite often that when a flat is not representing the vignetting correctly, that APP seems to correct in a wrong way, but this is then because the flat isn't actually presenting the vignetting properly. If your flat is good, then what you're left with should be the light pollution gradient.


ReplyQuote
(@jjones)
Brown Dwarf Customer
Joined: 8 months ago
Posts: 7
September 27, 2021 19:47  

So what's your take on that flat then?  I figured if it's there only in that Oiii and not in Ha or Sii then it would work to calibrate out that optic distortion or diagonal artifact?  or is there something else I'm missing or is that just maybe a defective Oiii filter?

 

I'm used to flats being much more smooth without that harsh line and it's showing in every result I take, but figured APP might have a way to calibrate that out since it was in the Flat?

 

I have used the light pollution piece heavily and it does take the gradients out, but not that section.  Speaking of LP, it wants you to select like 4 or 5 spots...  Does that mean select dark sky? or does that mean select light pollution areas?  I usually go around and select about 20-30 areas on the image with what I think are light pollution over dark sky areas with pretty good results, but kind of vague when I go use those tools 'What' I should be selecting... no nebula, some nebula, no star areas, dark background, pollution areas, etc?  if that makes sense.


ReplyQuote
(@jjones)
Brown Dwarf Customer
Joined: 8 months ago
Posts: 7
September 27, 2021 20:25  

So I did go back through NINA and created all new bias, dark, and then ran the full 'create master and darks' again and it created my new md-1, mf-1, and mb-1 and bad pixel map...

 

this time it seems to calibrate out perfect from what I can tell?  this is the same image but ran through with new files.  I really have no idea why my other didn't work.  I used the same flat file here, but the dark and bias were different?  I mean it seems way better...

lbn325 Oxygen III session 1 St

ReplyQuote
(@vincent-mod)
Universe Admin
Joined: 4 years ago
Posts: 3961
September 28, 2021 10:44  

Interesting, so that may point to something in the workflow perhaps? Maybe a mix-up of flats or something like that? Differences between filters is normal and can happen indeed, APP should be able to use that to correct the lights taken with that filter. If that doesn't work, usually there is an issue with the flat itself as it then doesn't match the light.


ReplyQuote
(@vincent-mod)
Universe Admin
Joined: 4 years ago
Posts: 3961
September 28, 2021 10:47  
Posted by: @jjones

I have used the light pollution piece heavily and it does take the gradients out, but not that section.  Speaking of LP, it wants you to select like 4 or 5 spots...  Does that mean select dark sky? or does that mean select light pollution areas?  I usually go around and select about 20-30 areas on the image with what I think are light pollution over dark sky areas with pretty good results, but kind of vague when I go use those tools 'What' I should be selecting... no nebula, some nebula, no star areas, dark background, pollution areas, etc?  if that makes sense.

The LP tool asks for a minimum of 5 boxes. You have to place these on the background where possible, for this I always stretch to the max (30% stretch preset) and increase the saturation to the max as well. This gives me a better idea as to where actual background is. The boxes don't have to be huge, just a few small ones on the background, if the data is challenging you slowly work your way up creating more small boxes towards the problem areas (not immediately on top) and check each time for the result.


ReplyQuote
(@wvreeven)
Quasar Admin
Joined: 3 years ago
Posts: 1226
September 28, 2021 12:29  
Posted by: @jjones

I mean it seems way better...

It may look that way but it doesn’t have to be. Can you save the NINA result in FITS format and load it as a light in APP? Then you can compare the background levels of both stacks which is the only way to really compare the two images. 


ReplyQuote
(@jjones)
Brown Dwarf Customer
Joined: 8 months ago
Posts: 7
September 28, 2021 16:30  

Thanks guys...  the information is most helpful...  So I think the issue may have been I was using stacked darks, bias from a while back, which I thought would be fine, but they were live stacked using an ASIair system.  NINA doesn't do stacks, so I can't really compare that with APP, but I did use NINA to retake a full set of Bias, Darks, and the Lights I reused because they were new and matched my data and that is how I ended up with latest good picture...

So, I proceeded to use my new bias and darks with my Ha flats and my Sii flats and restacked some older data and made a world of difference. So my conclusion was that the stacked bias/darks from my other live stacker in ASIair may not have worked with the Flats taken from NINA?  Not sure the difference, but images say differently.

I am going out to the observatory to retake a full set all from NINA of darks and just move on with it, but for anyone else struggling, only thing I can think of here was perhaps different meta data in the bias/darks and they weren't being applied?  perhaps different bit depth from a 12 bit camera to what APP wanted?  perhaps different temps or gains and they don't apply?  I'm now doing the bias and darks with a full set @ same temp & same gains on each setup. 


ReplyQuote
(@vincent-mod)
Universe Admin
Joined: 4 years ago
Posts: 3961
September 28, 2021 17:38  

Ah yes, APP only works with its own created calibration files. Other programs may not do that optimally or in the same way as APP does and that can cause issues.


ReplyQuote
(@jjones)
Brown Dwarf Customer
Joined: 8 months ago
Posts: 7
September 28, 2021 19:00  

@vincent-mod  It's the issue, and you can mark this resolved, it's working sooo much better with this, probably all my data for the last year didn't work.  I really would have thought my calibration files would work as they seem so similar... but great to know moving forward.  Here is my rosette stack of Oiii data and you can see the same stack with old calibration files not applying correctly and new one.

Rosette Oiii Oxygen III session 1 lpc St good calibration

 Good calibration

Rosette Oiii Oxygen III session 1 lpc St bad calibration

Bad calibration

Thanks!!  Sooo happy to have it working, was getting frustrated.

 

Last Question, and I've looked but not finding solid answer and would love to know this last little bit.

I understand SNR & Noise 0-1 normalized values...  I don't think these slight differences matter, but for explanation purposes

Good image SNR:  1.69E+00  (slightly better, I have been shooting to get over 5.00 for good signal images)
Bad image SNR:  1.58E+00

Good image Noise:  2.37E-04 (slightly lower, which my understanding is better)
Bad image Noise:  2.38E-04

 

What I am confused about, or maybe on the right track with is background and dispersion
Good Image Background:  1.025E-02  (Lower background signal is better? Equals darker background?)
Bad Image Background:  1.173E-02

Dispersion (sharpness of the signal?)
Good image dispersion:  2.40E-04 (Lower dispersion is better?)
Bad image dispersion:  2.61E-04 

 


ReplyQuote
(@vincent-mod)
Universe Admin
Joined: 4 years ago
Posts: 3961
September 29, 2021 12:03  

Regarding the numbers; lower background signal can mean it's better yes and lower dispersion means that, among other causes, the background level is more even across the sub.

Great it's working now! I always tell people to not wait too long to ask for assistance, it can greatly reduce frustration. 🙂


ReplyQuote
Share: