By legnolpi on Sunday, 03 October 2021
Posted in Syndicate Lounge
Replies 51
Likes 1
Views 1.2K
Votes 0
Hello all  i m strugglin since a few hours to process the file of job 1196.  i cant find a way to calibrate correctly Lum file . all the fits of this job are shoot in september so i use the latest calib file . Here is the result of the integration with WPBB with Lum file and calib file:  Lumintegrationwithflat.png We can see lots of dust donut  i try all sort of combination of calib file , corrected flat film by adding some value , it doesnt work  after a while i integrated without any flats :  lumintwithoutflat.png And here is my flat file used before :  LumFlat.png we can see that the big donut low-left , is not exactly at the same place on the flat file than on the integration file.  it is obvious if you switch rapidily between frames in pixinsight.  I think that the cam has rotate since the flat have been shoot or that dust have move .  As it is an open scope , i think that flats have to be taken more more often than one each 6months Also , with this camera , you should take way longer flat .  the LUM flat are 0.02s long exposure. I ve read that on some cmos camera , very short exposure raise problem. I think you should try a dimmer flat planel and have flats shoot with +1s long exposure .  i dont know if somebody have succeeded to deals with L file on this scope but if yes, please share the recipe   thansk  florent
Hi Florent,  Thanks for posting that. Like you I'd been trying a few things to at least reduce the effects of the dust but got no improvement apart from clipping the blacks and obliterating them and losing data. Put all the affected ones on hold for now.  In full agreement with you on this.  Regards,  Ray 
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Pete, is going to take new flats for this pier tomorrow. Its hard to put a time scale on flats, at times they last 2/3 months and other times its a few weeks A lot depends on the Spanish wind and season Cheers Steve
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Ok , thanks.  il will try them  
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Florent, How did you create your dark calibration files? I was getting the same effect you are seeing. I found that it went away if you create master darks without using a bias frame, and calibrate flats only with bias frame (no dark frame). Then use these masters in WBPP without supplying any bias frame. It means that you can't just tip everything into WBPP and let it sort it all out. You need to create the masters separately. I found this solution partly on Pixinsight forum, and partly somewhere else (can't remember where). I don't understand why it works, but (so far) it does for both piers 1 and 14.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
hi  do you have process this particular dataset with succes ?  I dont use BIAS  for calibrating i use   Light ,  Dark , Flat and DarkFlat  . If DarkFlat is very short , you can replace DarkFlat with Bias , it is nearly the same thing.  Bias is on all the file you get so if you   calibrated light frame with  Ligh - Dark - Bias , you are substrating two time the bias as  Dark =  DarkSignel + Bias That's why you dont need to put bias in WBPP but in some CMOS came , specially old one, very very short exposure (as in Lum Flat ) deliver strange result, the gain is not linear for short exposure so you can have strange result .  That why you need to get long ( > 1s) flat frame . 
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Florent, I will try my method today and let you know what happens.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Florent, I checked out Job 1186 with the flats method I have used successfully, and see exactly the same problem that you do. I tried various different ways of processing Flats and Darks, but always with the same result. I looked at the individual calibrated files and the dust doughnuts were always there, although some files were worse than others. I tried a different set of luminance files from job 1152 (Draco Group) that also were taken in September. Same problem. Apologies for going up a blind alley. There is definitely a problem with the current set of flats.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Nice you have test , it confirm i m not nuts
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
I tested with multiple methods also, the donuts are always there. Not just for this project, the NGC488 also has the same issue, albeit less noticeable.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
I tested with multiple methods also, the donuts are always there. Not just for this project, the NGC488 also has the same issue, albeit less noticeable.
Looking at the dates of the individual subs for NGC488, it looks like some were taken before the maintenance visit and some after. We would need to make sure the right flats are used.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
I tested with multiple methods also, the donuts are always there. Not just for this project, the NGC488 also has the same issue, albeit less noticeable.
Looking at the dates of the individual subs for NGC488, it looks like some were taken before the maintenance visit and some after. We would need to make sure the right flats are used.
Good spot I did say at the time when all the new data was taken from after the maintenance visit. So this may indeed be the issue as we had to get rid of the internal reflection so the filters were 100% removed on the last trip and cleaned Steve
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Do you have time to make new flats ?  florent
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Yes not a problem but the skies have been clear this week so we dont waste imaging time collection flats As soon as we know the roof is closed for cloud we will be getting new ones, however did you check the dates on the fits header just in case ? Steve
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
job 1186 was shoot after maintenance ,  frame between 19/09 to 28/09 
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
job 1186 was shoot after maintenance ,  frame between 19/09 to 28/09 
Cool, as soon as the next cloudy night comes along we will take some new flats :0 Steve
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Thanks Steve.
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
Pete has just completed new Flats, I am sure Phil will let you all know when available
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
yeah , nice news
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
I processed the new flats and soem new datasets this weekend. Stil there are 3D donust from the dust, especially for the Lum filter. t seems that the dust patterns in the MasterFlat and light frames are not aligned. I then integrated all Lum from NGC 772 without registering to highlight the dust patterns and compared with the MasterFlat of Oct in pixinsight. Sure enough, some of the dust clearly shifted, by around 10-20 pixels. The new flats were taken from Oct 13 nad the Lum frames of NGC 772 were mostly in Oct 12 so I don't think think likely the dusts changed during this time.  Do you guys have any ideas why this happened? Could there be shifts/sags in the system or the angle of lights hitting the sensor?
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
i ve process the job i 1186 NGC5907, still the donut .  the dust is not in the same place on flat frame or lum flat, i used the news one  ; oct. 
·
2 years ago
·
1 Likes
·
0 Votes
·
0 Comments
·
View Full Post