Calibration Problem

[Starlight]

Hi
I hope someone can help me
I've just started using Pixinsight 1.7 and am having a bit of a problem
When I tried to calibrate images of NGC6888 with Darks, Bias & Flat frames the resulting image only had the very brightest stars visible even when viewed with Screen Transfer Function. See the 4th image attached below.
I experimented on a single image and found that if I just used one uncalibrated flat all of the stars and nebula were still visible – although vignetting worse than for the uncalibrated light –and if I used a single calibrated flat only the brightest stars were visible and vignetting even worse – all of the above when viewed with STF.
I've attached a screen shot to show you what I mean.

Top left is a single uncalibrated light
Top right is a light with an uncalibrated flat
Bottom right is an uncalibrated flat
Bottom right is a calibrated flat

I tried again using the Batch Preprocessing script and it terminated with an error message as below.

When I looked at the output images they were also generally as for my single image attempt.

Thanks in anticipation
Don

ps

[sreilly]

Are you really using a scaled dark? If your dark is the same time and temperature then forget the bias frame.

-Steve

[sreilly]

You may want to try just using the Image Calibration tool if you are using the Batch Processing script now. Just to see if there is a difference.

-Steve

[Starlight]

Thanks Steve
The Darks were 3 minutes as original lights and weres taken immediately after completion of light series
I've tried calibration again using a single image and four different combinations
Used Bias, Dark & Flat
Used Bias only
Used Dark only
Used Flat only

See Ist image for settings and second for results.

Image calibrated with Bias, Dark & Flat only showed brightest stars
Image calibrated with Bias was very similar to original uncalibrated image
Image calibrated with Dark only showed brightest stars
Image calibrated with Flat was overcorrected for vignetting compared with original

I expect I'm doing something simple wrong with the settings but can't figure out what.

Don

[sreilly]

I assume you are using a dSLR camera. Have you used the Debayer process on the images? I'm not familiar with dSLR imaging so I can't speak to those processes. You also need to process your flats and darks. Have you looked at this tutorial http://www.pixinsight.com/tutorials/master-frames/en.html ? Have you seen this? http://pixinsight.com/forum/index.php?topic=2570.0

[tsaban]

Hi Don,
I think I had the same problem as you describe (btw. I shot with a DSLR). I am still not 100% sure but I suspect the reason is the preprocessing script tries to calibrate the falts with the dark frame. In my case the exposure lengths are very different (5min vs 1/4sec) and the dark scaling goes wrong. When I calibrate my images manually (please search the forum on how to do it) I get very good results.

Hope this helps
Tahir

[pfile]

yeah, that's right - i'm reasonably sure everything (lights, flats) in BatchPreprocessing gets calibrated with the darks or master dark.

this is probably not appropriate for the flats... so what i've been doing over the last couple of days is making my own master flats which have been integrated separately with a matching master bias. i shot my flats at iso100 (vs 1600 for the lights), so i have to do this.

i can't tell if the script is hardcoded to output monochrome CFA master calibration frames, or if i accidentally told it to do that. but my master dark and master bias (iso1600) were in mono CFA format, so i had to make my new master flat like that. so i configured DSLR_RAW for mono CFA (is there a format hint that does this? "raw" gives me an RGB bayered image), then made my master flat bias, calibrated the flat subs with it, and then integrated the flat subs into a master flat.

then back in BPP i select the 3 masters and tell it to use masters rather than to create them. i think this is "safe", meaning the script does not try to calibrate the master frames (i hope).

the BPP script is a godsend but it does need some more improvement, especially in this department. even if you are using a mono CCD it's likely your flat exposures are short and need a different dark. and if you are using a DSLR and use low iso for the flats, same situation...

now of course for some reason the debayering in the BPP script is choking on my calibrated lights (which are now mono CFA). they are fine though - they can be debayered with the Debayer process, and then registered manually and integrated. i'm not sure what's causing this. i tried unchecking "CFA" but then my dark scaling went wild (huge values.) that's usually an indication that something went wrong with the debayer process, so i think it needs to be checked.

[Starlight]

Thanks guys
Following your suggestions I fiddled around with the settings and agree that the problem is with the Dark optimise check box.
If I leave it unchecked everything works much better - see attachment of result with single file of each.

Don
ps I was using a QHY8 - OSC

[Cleon_Wells]

Hi Don, question, how many uncalibrated dark subs did you integrate for the Master Dark?
Cleon

[Starlight]

Cleon
In the last screen shot there was only one as I wanted to try a quick test with one dark, bias, flat and light.
Originally I used 3x3 minutes - I know thats not really enough but they often say that the QHY8 doesn't really need any as I don't get very many hot pixels.

Don

[Cleon_Wells]

Don, for the hell of it try taking at least 20x3min dark subs right now and convert them to 32bit fits files, then integrate them with out pixel rejection to create your un_calibrated Master Dark. I think this MasterDark will work with both Cal and Opt checked. I have been using  a Canon T1i camera and with only 120sec Darks at iso 400 and I have had problems with my MasterDark file if I tried subtracting the bias pedestal from each dark sub , ( with the resulting cal dark subs having a number of pixels without any signal , zero level ) , before I integrated them .
Cleon