Share:
Notifications
Clear all

[Solved] Flat calibration


(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

Hello!

I have you used APP about 1,5 years and been satisfied many times with results. But sometimes I don't understand why something went wrong. One of these moments is now.

 

I tested first time my ASI294MM with EdgeHD800 and OAG. I took all the calibration frames (incl. darks, flats, dark flats and bias frames) in normal way as I did before. But I got very illogical results. There is short integration samples what went wrong.

Integration 3 x 900s (Oiii 3nm) WITH flats:

With flat

Integration 3 x 900s (Oiii 3nm) WITHOUT flats:

No flat

Master flat:

Master flat

So when I use flats there is more uneven background than without. I think that something went wrong, but have you any advices what? I want to highlight that data is collected using flatbox, with right methods and obeying calibration rules. As in many more successful cases before. Please, could you help me understand what happened?

Best regards,

Tommi Liinalampi


ReplyQuote
(@wvreeven)
Galaxy Admin
Joined: 4 years ago
Posts: 2035
 

@tojuliin Hi Tommi,

Have you been able to take good flats with the same OIII filter before? Can you post the histogram of a single flat and of the master flat?

Wouter


ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

@wvreeven

Hello!

I have used this filter before and there were also same problem with Ha and Sii frames. I haven't had problems with filters and F-wheel before when I have used WO 81 APO. Example this picture calibrated successfully with APP (3 week ago):

NGC7822

But you asked histogram of this new issue...

Histogram of Master flat, stretched image:

image

And linear image (maybe more relevant?):

image

And ideas?

This post was modified 2 months ago by Tommi Liinalampi

ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

And you asked also histogram of just one flat frame. 

(linear, no stretched)

image

 


ReplyQuote
(@wvreeven)
Galaxy Admin
Joined: 4 years ago
Posts: 2035
 
Posted by: @tojuliin

And ideas?

Unfortunately not.Would you mind uploading a set of 5 lights and 5 of each raw calibration frame so I can have a look and try a few things myself? Upload instructions can be found at the top of any forum page. Please create a folder called tojuliin_oiii and put the files in there. Let me know when the upload has finished and I'll have a look as soon as I can.


ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

@wvreeven 

Hello!

I uploaded the files you asked. Only 3 light frames, because I haven't more (yet). I took new dark, dark flat and bias -frames last night and I took these very carefully that I'm sure that there were no light leaks. But nothing changed.

I didn't mentioned yesterday that I got critical warning in one calibration attempt. I got this again when I tried to calibrate with new dark- and bias-frames. It was this before integration:

Sieppaa

I'm not sure what is the specific reason to that. My data is correct. Only thing which could problem for APP is the fact that I take images in bortle 3, using very short narrowband filter (3nm) and with quite slow telescope (EdgeHD F/7). I think I haven't possibility to take longer exposures than 15 minutes with my setup. But this should not be a problem for program I think.

Hope you find the problem using uploaded frames.

 

Thank you for help,

Tommi


ReplyQuote
(@wvreeven)
Galaxy Admin
Joined: 4 years ago
Posts: 2035
 
Posted by: @tojuliin

Only thing which could problem for APP is the fact that I take images in bortle 3, using very short narrowband filter (3nm) and with quite slow telescope (EdgeHD F/7). I think I haven't possibility to take longer exposures than 15 minutes with my setup. But this should not be a problem for program I think.

The popup is shown when pixels in the lights clip after applying the calibration masters. If you are sure that there is no light leak then this indeed probably is because of a (slightly) too short exposure time. F/7 is quite slow indeed. With my F/5 telescopes I need to expose for longer than 5 but shorter than 10 minutes to not get the popup.

I'll have a look at the files you uploaded to see if I can see what's causing the flat over correction.


ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

@wvreeven 

I noticed that the background is quite dark. I have looked black point number in APP sidebar when using default stretch option and in this case it is just little more than 1600, but when I have used my APO F/5.5 it's often more than 1900 using 600 s subs. SC F/7 telescope is about 1,6 times slower than F/5.5, so I have used 900 s exposure time (1,5 times longer than with APO). But 1600 is less than 1900 if we look black point. 

image

But there are many difficulties if I increase my exposure time. One of these is tracking. iOptron CEM70 is quite good, but not perfect for very long exposures with long focal length. And dithering is also a problem. Project should be very big if you want enough of different positions using dithering. Longer exposure time is not so favorite solution indeed. I could increase gain a little bit if it helps, but I'm not sure about that?

 

Please, check the files, if there is something to do or some method to calibrate these. I'm lucky because I noticed this issue so early that I have just only 3 x Ha, Oiii, Sii taken.

 

Tommi


ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

Now I tried with different kind of data and APP did same. I tried to calibrate my images of M74 but after using flat background was more uneven than without. This was LRGB data and 180 s (gain 120). I think that exposure time should be okey. Black point in luminance date was about 1850 and in color data was about 1650. I took this data in same night and flats on the next day like data I uploaded before.

This post was modified 2 months ago 4 times by Tommi Liinalampi

ReplyQuote
(@stastro)
Red Giant Customer
Joined: 3 years ago
Posts: 124
 

Why are you taking Dark Flats and BIAS frames?

With CMOS Cameras you should only be taking Dark Flats, BIAS frames are useless on modern CMOS cameras, but either way you should not be using both BIAS and Dark Flats with any camera, it's one or the other.

I only use Dark Flats because I am using a CMOS Camera, when I used a CCD Camera, I would only use BIAS frames


ReplyQuote
(@wvreeven)
Galaxy Admin
Joined: 4 years ago
Posts: 2035
 

@stastro APP needs bias to be able to subtract the dark current pedestal from the calibration files.


ReplyQuote
(@stastro)
Red Giant Customer
Joined: 3 years ago
Posts: 124
 
Posted by: @wvreeven

@stastro APP needs bias to be able to subtract the dark current pedestal from the calibration files.

That's the whole point in Dark Flats though.  It is pretty common knowledge that CMOS Cameras do not work well performing 0 second exposures to obtain BIAS signal.  BIAS works well on CCD, but is totally flawed on modern CMOS cameras.  There are so many threads on this.

Dark flats contain the "BIAS" or "Dark Current" or "Electrical Signature" that would be used to calibrate the Flats, so you should never have both BIAS and Dark Flats in a calibration, as you are going to make things abundantly worse for sure.

 


ReplyQuote
(@tojuliin)
White Dwarf Customer
Joined: 2 years ago
Posts: 21
Topic starter  

@wvreeven 

Hello!

Now it seems that my light files could be taken with WRONG offset rate, which shouldn't be possible using Asiair(!). I consulted finnish astrophotography community and we tried to find answer why my light files were even darker than dark files. You can't even change offset rate in Asiair, but maybe there was some program error or something and that's why offset wasn't 30 as in calibration files. We thought that there couldn't be any other answers to that. Even in image headers included wrong information that offset was 30. It was less!

I think that there is no problem with APP calibration process. Or the result was same like example using Pixinsigt, which also failed same way in calibration process.

Of course, if you have any ideas how to calibrate these with APP I would be happy. But now I would like to thank you for support.

Best regards,

Tommi


ReplyQuote
(@wvreeven)
Galaxy Admin
Joined: 4 years ago
Posts: 2035
 

@stastro You are right about modern CMOS cameras not doing well when taking 0 sec (or rather 0.001 sec) bias frames, which is why we recommend taking 0.1 sec bias, which still is good enough. See also https://www.astropixelprocessor.com/community/main-forum/dark-flats-and-or-bias-frames-and-darks/#post-25266


ReplyQuote
(@stastro)
Red Giant Customer
Joined: 3 years ago
Posts: 124
 

@wvreeven The challenge you have is most modern CMOS cameras are really bad at taking very short exposures, hence why so many people have issues with flats that are sub 2 seconds.  But again, this is the whole purpose of dark flats, if you use dark flats, you should not use BIAS frames

https://www.astrobin.com/forum/c/astrophotography/deep-sky-processing-techniques/bias-versus-dark-flat-frames/

 


ReplyQuote
Share: