Splotchy background after Batch Preprocessing and DBE or ABE

[pfile]

here is what i see. the top left is the raw light you posted, uncalibrated. the top right is debayered using RGGB and viewed under STF with channels unlocked. below that is the _c_d file you posted viewed under STF with channels unlocked. with channels locked it would look green. the _c_d file should look like the top right file, but for some reason it does not.

i believe that the splotchyness is a byproduct of this problem and it needs to get solved before worrying about anything else.

rob

[sharkmelley]

my bias, darks, and lights were created in SGP.

Are you sure?
If so, why does the dark frame have the "starburst" amp glow on the left hand side but in the light frame it is on the right hand side?  What I'm saying is that your example dark is incompatible with your example light even though you say both were created in SGP.

Your example flat is also incompatible with your example light.

This is why we are seeing that grid/chequerboard/screen door problem from the calibrated files, once they are debayered.  It is not normal to see this pattern in a debayered file.

Mark

[alanrock]

Let me clarify. The images I just posted (checkerboard and screen) were captured from pix produced while using darks, bias and lights created with SGP and flats created with Maxim, although the files I had posted earlier used darks which were created with Maxim. Also, the darks I used this last time were 300 sec @ -10º instead of 300 sec at -15º (because as I went through my darks, I saw that these were the ones I had created with SGP)--I just wanted to see if that helped, but I'm guessing that until I get flats created with SGP I'm going to have this same problem. Make sense?

[pfile]

i think so. but what happens if you just run without flats? if the debayered files look OK (like my 2nd image in the screenshot) then you'll know it's the flats that are causing the problem for sure.

rob

[alanrock]

Will BPP run without the flats?

[alanrock]

OK, I think you folks have figured it out for me. I just ran BPP without flats, but WITH Bias, Darks and Lights all created in SGP, and the screen door effect is gone! Here's the final result with no flats applied: https://drive.google.com/open?id=1X6dinyZxcoxHjm0_5AJ_XEhhVFfKsPui
The image looks quite correct I think (other than the fact that it needs flats!). What do you think?

[pfile]

yes that looks proper.

on the subject of the darks, i think when you have amp glows you need to avoid dark scaling (optimization). so for this sensor it would be best to make darks that match the lights in duration, and then calibrate without bias frames making sure "optimize" is unticked.

rob

[alanrock]

Thanks. I didn’t actually have the optimize button checked, but I also didn’t have 600 sec darks because I’d never done 600 sec lights before, I’ll have to build up my library. Are you saying that I don’t need to use bias frames if I’m using appropriate darks with this camera? Thanks again for all your help!!

[sharkmelley]

Thanks. I didn’t actually have the optimize button checked, but I also didn’t have 600 sec darks because I’d never done 600 sec lights before, I’ll have to build up my library. Are you saying that I don’t need to use bias frames if I’m using appropriate darks with this camera? Thanks again for all your help!!

Looking at the example files you provided, the light and dark both have the "starburst" amp glow but the bias did not.  So to calibrate out the starburst you need closely matched lights and darks (i.e. same length exposure and same temperature) and you must disable dark optimisation (which you have done).  But you still need bias frames in order for the flats to work correctly.

Mark

[sharkmelley]

Thanks. I didn’t actually have the optimize button checked, but I also didn’t have 600 sec darks because I’d never done 600 sec lights before, I’ll have to build up my library. Are you saying that I don’t need to use bias frames if I’m using appropriate darks with this camera? Thanks again for all your help!!

Looking at the example files you provided, the light and dark both have the "starburst" amp glow but the bias did not.  So to calibrate out the starburst you need closely matched lights and darks (i.e. same length exposure, same temperature, same offset and gain) and you must disable dark optimisation (which you have done).  But you still need bias frames in order for the flats to work correctly.

Mark

[alanrock]

So this probably isn’t a PI question, but does anyone know of any way to make my flats usable?

[sharkmelley]

So this probably isn’t a PI question, but does anyone know of any way to make my flats usable?

Run the FastRotation process on them - either Horizontal Mirror or Vertical Mirror or both.

Mark

[alanrock]

Thanks! I guess I’ll try horizontal, then BPP again and if that doesn’t work I’ll run the vertical mirror and try again.

[pfile]

you should be able to tell what transform you need for sure by loading one of your lights with the fits reader set to up-bottom and then loading the same light set to bottom-up and see what transform you need to make them match.

but dont forget to set the reader back to whatever it was set to when you are done!

rob

[alanrock]

Well, I did a vertical mirror based on the advice above, but for some reason when I save the file and re-open it, there's a new file with an _thumbnail suffix appended. When I tried running BPP with the newly saved files, it blew up, I believe because of the thumbnails, which incidentally don't show up as separate files.
I got this:

Calibrating target frame 1 of 28
[2018-08-21 21:31:30] Loading target frame:
[2018-08-21 21:31:30] /Volumes/Astronomy Pix/Pictures/Astronomy/ZWO_Flats/Flats-0001 copy.fit
[2018-08-21 21:31:30] * Subimage 1 of 2
[2018-08-21 21:31:30] Reading FITS image: 16-bit integers, 1 channel(s), 5496x3672 pixels: done
[2018-08-21 21:31:31] 46 FITS keywords extracted.
[2018-08-21 21:31:31] * Subimage 2 of 2
[2018-08-21 21:31:31] Reading FITS image: 8-bit integers, 1 channel(s), 400x267 pixels: done
[2018-08-21 21:31:31] *** Error: Incompatible image geometry